# Vitess v23.0.0

2025-11-04 · [Original release notes](https://github.com/vitessio/vitess/releases/tag/v23.0.0) · [Full summary](https://ratatosk.io/en/releases/vitess/v23.0.0)

Security 11 · Breaking changes 27 · Deprecations 15 · Important fixes 49 · Enhancements 47 · Dependency updates 17 · Other changes 110

## Check before upgrading

### Prerequisites

- **Manual sequence for `vitess-operator` upgrades from `8.0.x` to `8.4.x`**: Add `innodb_fast_shutdown=0` in the YAML extra cnf and apply it, wait for all pods to be healthy, change the image to `vitess/lite:v23.0.0`, then remove the setting and apply again. This sequence is needed only for the move to `8.4.x`, not for upgrades or downgrades between `8.4.x` versions.
  > >  ⚠️ Upgrading to this release with vitess-operator:
  > >
  > > If you are using the `vitess-operator`, considering that we are bumping the MySQL version from `8.0.40` to `8.4.6`, you will have to manually upgrade:
  > >
  > > 1. Add `innodb_fast_shutdown=0` to your extra cnf in your YAML file.
  > > 2. Apply this file.
  > > 3. Wait for all the pods to be healthy.
  > > 4. Then change your YAML file to use the new Docker Images (`vitess/lite:v23.0.0`).
  > > 5. Remove `innodb_fast_shutdown=0` from your extra cnf in your YAML file.
  > > 6. Apply this file.
  > >
  > > This is only needed once when going from the latest `8.0.x` to `8.4.x`. Once you're on `8.4.x`, it is possible to upgrade and downgrade between `8.4.x` versions without needing to run `innodb_fast_shutdown=0`.

### Known issues

- **Experimental throttler remains under active development**: Work is ongoing, and breaking changes may occur in future releases.
  > Vitess v23.0.0 introduces a new, experimental Query Throttler framework for rate-limiting incoming queries ([RFC issue #18412](https://github.com/vitessio/vitess/issues/18412), [PR #18449](https://github.com/vitessio/vitess/pull/18449), [PR #18657](https://github.com/vitessio/vitess/pull/18657)). Work on this new throttler is ongoing with the potential for breaking changes in the future.
- **Availability risks from disabling `EmergencyReparentShard`-based recoveries**: Frequent reliance on this control may indicate an anti-pattern.
  > **Note: disabling `EmergencyReparentShard`-based recoveries introduces availability risks; please use with extreme caution! If you rely on this functionality often, for example in automation, this may be signs of an anti-pattern. If so, please open an issue to discuss supporting your use case natively in VTOrc.**

## Security (11)

### Consul static-file authentication requires a loaded credential

- Affects: setups using `--consul-auth-static-file`
- Additional action: The JSON file must allow at least one credential to be loaded.

Previously, an empty or invalid JSON file—including `{}`—could let startup continue without authentication. Such files now cause startup to fail.

> ### <a id="minor-changes-topo"/>Topology</a>
> #### <a id="consul-auth-static-file-check-creds"/>`--consul-auth-static-file` requires 1 or more credentials</a>
> The `--consul-auth-static-file` flag used in several components now requires that 1 or more credentials can be loaded from the provided json file ([PR #18152](https://github.com/vitessio/vitess/pull/18152)).
> **Impact**: Configurations with empty or invalid credential files will now fail at startup rather than silently continuing with no authentication.
>  * Check `consultopo` loaded `>=1` credentials w/`-consul_auth_static_file` [#18152](https://github.com/vitessio/vitess/pull/18152)

### GetSchema respects the `EnableViews` setting

- Affects: tablet servers with `EnableViews` disabled
- Additional action: `EnableViews` disabled

Previously, clients could still fetch view definitions through `GetSchema` despite that setting. Now, `SchemaTableType_VIEWS` returns an empty result, and `SchemaTableType_ALL` returns only table definitions.

>  * Fix GetSchema RPC to prevent returning view definitions when EnableViews is disabled [#18513](https://github.com/vitessio/vitess/pull/18513)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `vtctldclient GetPermissions` clears password checksums

- Affects: setups using `vtctldclient GetPermissions`

The JSON output omits unpopulated fields.

>  * [release-23.0] Potential fix for code scanning alert no. 2992: Clear-text logging of sensitive information (#18754) [#18760](https://github.com/vitessio/vitess/pull/18760)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `vtctldclient GetPermissions` response omits `authentication_string`

- Affects: setups that use `vtctldclient GetPermissions`

The patch also notes that the field is redacted server-side in v23 and later.

>  * [release-23.0] `vtctldclient GetPermissions`: hide `authentication_string` from response (#18771) [#18799](https://github.com/vitessio/vitess/pull/18799)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### External-tablet VReplication permission check no longer queries `mysql.user`

- Affects: external tablets with reverse replication enabled
- Additional action: The permission check runs during a write-traffic switch when reverse replication is enabled.

Previously, the permission check queried `mysql.user`, requiring access to that table and making checks difficult to test reliably across wildcard grants, role-based authentication, and database versions. During a write-traffic switch, it now tests permissions with dummy DMLs in a transaction without requiring access to `mysql.user`.

>  * VReplication: permission check logic on external tablets -- switch to a more practical solution [#18580](https://github.com/vitessio/vitess/pull/18580)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### [Security status unclear] VTAdmin TLS support for vtctld gRPC connections

- Affects: VTAdmin setups using TLS for the vtctld gRPC connection
- Additional action: Configure `--vtctld-grpc-ca`, `--vtctld-grpc-cert`, `--vtctld-grpc-key`, `--vtctld-grpc-server-name`, and `--vtctld-grpc-crl` for vtctld gRPC TLS.

Previously, TLS-only vtctld connections prevented VTAdmin from starting because its proxy used insecure gRPC credentials. With secure dial configuration available, VTAdmin can connect to vtctld over TLS.

>  * Add vtctldclient TLS support to vtadmin GRPC [#18556](https://github.com/vitessio/vitess/pull/18556)
> Five new TLS-related flags were added for secure vtctld connections ([PR #18556](https://github.com/vitessio/vitess/pull/18556)):
> | Flag | Description |
> |:-----|:------------|
> | `--vtctld-grpc-ca` | CA certificate file for vtctld gRPC TLS |
> | `--vtctld-grpc-cert` | Client certificate file for vtctld gRPC TLS |
> | `--vtctld-grpc-key` | Client key file for vtctld gRPC TLS |
> | `--vtctld-grpc-server-name` | Server name for vtctld gRPC TLS validation |
> | `--vtctld-grpc-crl` | Certificate revocation list for vtctld gRPC TLS |
> These flags enable mTLS (mutual TLS) authentication between VTAdmin and vtctld for enhanced security.

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### File backup path traversal checks added

- Affects: installations using file backup storage (Inferred)

Joined paths that traverse outside the configured storage root now return an error instead of being used. This check applies to file-backup read, list, start, and remove operations.

> ### Security 
> #### Backup and Restore
>  * [release-23.0] Address dir traversal in file backup storage `GetBackups` RPC (#18814) [#18818](https://github.com/vitessio/vitess/pull/18818) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Java JDBC driver vulnerability fixed by replacing `commons-lang`

- Affects: setups using the Vitess Java JDBC driver (Inferred)

The driver previously used `commons-lang` 2.6 and imported `org.apache.commons.lang.StringUtils`; after the upgrade, it uses `commons-lang3` 3.19.0 and imports `org.apache.commons.lang3.StringUtils`.

> #### Java
>  * [release-23.0] Resolve `commons-lang` vulnerability in Java driver (#18768) [#18797](https://github.com/vitessio/vitess/pull/18797) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTAdmin JavaScript dependencies updated

- Affects: setups using VTAdmin

The updates are intended to address Dependabot vulnerabilities in VTAdmin’s JavaScript dependencies.

> #### VTAdmin
>  * [release-23.0] Address `Moderate` dependabot vulns in VTAdmin (#18744) [#18750](https://github.com/vitessio/vitess/pull/18750)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTAdmin dependency updates address Dependabot alerts

- Affects: setups using VTAdmin (Inferred)

VTAdmin dependencies had triggered Dependabot alerts. The package lockfile now includes updates intended to resolve them, along with updates to other packages.

>  * [release-23.0] Address `Moderate` dependabot vulns in VTAdmin (#18744) [#18750](https://github.com/vitessio/vitess/pull/18750)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTAdmin Vite dependencies updated for security

- Affects: setups using VTAdmin (Inferred)

The previous dependency ranges had Dependabot security alerts. The manifest and lockfile now use `vite` `^7.1.12`, `@vitejs/plugin-react` `^5.1.0`, `vite-plugin-svgr` `^4.5.0`, and `vitest` `^4.0.3`.

>  * [release-23.0] vtadmin: upgrade vite to the latest (#18803) [#18812](https://github.com/vitessio/vitess/pull/18812) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

## Breaking changes (27)

### Four deprecated VTGate metrics removed

- Affects: VTGate monitoring dashboards or alerting systems using any of the four removed metrics

The metrics had been deprecated in v22.0.0 and are no longer provided. `QueryExecutions` replaces `QueriesProcessed`, `QueryRoutes` replaces `QueriesRouted`, and `QueryExecutionsByTable` replaces both by-table metrics.

> Four deprecated VTGate metrics have been completely removed in v23.0.0. These metrics were deprecated in v22.0.0:
> | Metric Name               | Component | Deprecated In |
> |---------------------------|-----------|---------------|
> | `QueriesProcessed`        | vtgate    | v22.0.0       |
> | `QueriesRouted`           | vtgate    | v22.0.0       |
> | `QueriesProcessedByTable` | vtgate    | v22.0.0       |
> | `QueriesRoutedByTable`    | vtgate    | v22.0.0       |
> **Impact**: Any monitoring dashboards or alerting systems using these metrics must be updated to use the replacement metrics introduced in v22.0.0:
> - Use `QueryExecutions` instead of `QueriesProcessed`
> - Use `QueryRoutes` instead of `QueriesRouted`
> - Use `QueryExecutionsByTable` instead of `QueriesProcessedByTable` and `QueriesRoutedByTable`
> See the [v22.0.0 release notes](https://github.com/vitessio/vitess/blob/main/changelog/22.0/22.0.0/release_notes.md#new-vtgate-metrics) for details on the new metrics.
> | Component |        Metric Name        | Was Deprecated In |                     Deprecation PR                      |
> |:---------:|:-------------------------:|:-----------------:|:-------------------------------------------------------:|
> | `vtgate`  |    `QueriesProcessed`     |     `v22.0.0`     | [#17727](https://github.com/vitessio/vitess/pull/17727) |
> | `vtgate`  |      `QueriesRouted`      |     `v22.0.0`     | [#17727](https://github.com/vitessio/vitess/pull/17727) |
> | `vtgate`  | `QueriesProcessedByTable` |     `v22.0.0`     | [#17727](https://github.com/vitessio/vitess/pull/17727) |
> | `vtgate`  |  `QueriesRoutedByTable`   |     `v22.0.0`     | [#17727](https://github.com/vitessio/vitess/pull/17727) |

### Default MySQL version for `vitess/lite:latest` changes to `8.4.6`

- Affects: the `vitess/lite:latest` image
- Action required: For `vitess-operator`, the one-time `8.0.x` to `8.4.x` sequence: add `innodb_fast_shutdown=0` to the YAML extra cnf and apply it; wait for all pods to be healthy; change the YAML to use the new Docker images; then remove the setting and apply the YAML again.

The tag previously selected MySQL `8.0.40`; after upgrading, it selects `8.4.6` without a change to the tag.

> The default major MySQL version used by our `vitess/lite:latest` image is going from `8.0.40` to `8.4.6`.
> This change was merged in [#18569](https://github.com/vitessio/vitess/pull/18569).
>  * Update MySQL 8.4 to be the default [#18569](https://github.com/vitessio/vitess/pull/18569)
> #### Online DDL
>  * Online DDL: support `alter vitess_migration ... postpone complete` syntax [#18118](https://github.com/vitessio/vitess/pull/18118)

### VTGate now advertises MySQL version `8.4.6` by default

- Affects: VTGate deployments
- Action required: If the desired advertised MySQL version is not `8.4.6`, set the `mysql_server_version` flag to that version.

Previously, VTGate advertised MySQL version `8.0.40`; the new default may differ from the MySQL version you are running.

> VTGate also advertises MySQL version `8.4.6` by default instead of `8.0.40`. If that is not what you are running, you can set the `mysql_server_version` flag to advertise the desired version.
>  * Update MySQL 8.4 to be the default [#18569](https://github.com/vitessio/vitess/pull/18569)

### VTGate `QueryExecutionsByTable` counts successful queries only

- Affects: setups that collect or use the VTGate `QueryExecutionsByTable` metric (Inferred)

Previously, failed query attempts were included, so the metric could overstate successful query volume. Failed queries are now tracked separately by error metrics, and values may decrease for workloads with significant failures.

> #### VTGate QueryExecutionsByTable Behavior Change
> The `QueryExecutionsByTable` metric now only counts successful query executions ([PR #18584](https://github.com/vitessio/vitess/pull/18584)). Previously, it counted all query attempts regardless of success/failure.
> **Impact**:
> - Metric values may decrease if your workload had significant query failures
> - More accurate representation of successful query volume
> - Failed queries are tracked separately via error metrics
>  * fix: emit QueryExecutionsByTable for success cases only [#18584](https://github.com/vitessio/vitess/pull/18584) 

### VTTablet `--skip-user-metrics` flag

- Affects: VTTablet setups emitting user-based metrics

In environments with many unique users, per-user username labels can cause metric cardinality explosion. When `--skip-user-metrics` is enabled, the username label is replaced with `UserLabelDisabled`.

> - `--skip-user-metrics` flag if enabled, replaces the username label with "UserLabelDisabled" to prevent metric explosion in environments with many unique users ([PR #18085](https://github.com/vitessio/vitess/pull/18085)).
> | `--skip-user-metrics` | vttablet | If enabled, replaces the username label in user-based metrics with "UserLabelDisabled" to prevent metric cardinality explosion in environments with many unique users. | [#18085](https://github.com/vitessio/vitess/pull/18085) |
>  * [vttablet] Emit Constant “userLabelDisabled” When skip-user-metrics is Enabled [#18085](https://github.com/vitessio/vitess/pull/18085)

### Vitess Bullseye Docker images no longer built or published

- Affects: Vitess Docker images based on `debian:bullseye`

Builds continue for Debian Bookworm and add Debian Trixie; the default Debian tag remains unchanged.

> [Bullseye went EOL 1 year ago](https://www.debian.org/releases/), so starting from v23, we will no longer build or publish images based on debian:bullseye ([PR #18609](https://github.com/vitessio/vitess/pull/18609)).
> Builds will continue for Debian Bookworm, and add the recently released Debian Trixie. v23 explicitly does not change the default Debian tag to Trixie.
>  * docker: add trixie and rm bullseye from build matrix [#18609](https://github.com/vitessio/vitess/pull/18609)

### Datadog tracer debug mode now defaults to off

- Affects: installations using the Datadog tracing plugin (Inferred)

Previously, the plugin always enabled debug mode, producing extremely verbose logging that could reach millions of logs depending on traffic. The `--datadog-trace-debug-mode` option turns it on.

Evidence: “Previously debug mode was hardcoded to true for the tracer in the datadog plugin, which made logging extremely verbose. This can easily add up to millions of logs depending on the amount of traffic your application receives, which can be very costly.” https://github.com/vitessio/vitess/pull/18347

>  * Configure datadog trace debug mode [#18347](https://github.com/vitessio/vitess/pull/18347)
> | `--datadog-trace-debug-mode` | All components | Makes Datadog trace debug mode configurable instead of always-on. Default: false. | [#18347](https://github.com/vitessio/vitess/pull/18347) |

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Immediate Online DDL cut-over for positive `--force-cut-over-after` values of `<= 1ms`

- Affects: Online DDL with a positive `--force-cut-over-after` value of `<= 1ms`

Previously, forced cut-over waited until `sinceReadyToComplete` exceeded the configured value. Now cut-over can happen even when the measured elapsed time is lower.

Evidence: “- When `--force-cut-over-after` value is `<= 1ms` we consider it as "immediate" even if we somehow measure the time-since-ready to be less than that.” https://github.com/vitessio/vitess/pull/18423

>  * Online DDL cutover enhancements [#18423](https://github.com/vitessio/vitess/pull/18423)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTOrc restarts direct replicas for unreachable primaries

- Affects: VTOrc-managed replication

For `UnreachablePrimary`, VTOrc stops and starts replication on one arbitrary direct replica. If only some replicas are failing to replicate, it classifies the condition as `UnreachablePrimaryWithBrokenReplicas` and restarts all direct replicas; restarts are limited to one per minute per tablet.

Evidence: “However, if there is a non-empty subset of replicas that are failing to replicate (and these are **not** all of the replicas) then we analyze as a new `UnreachablePrimaryWithBrokenReplicas`. This scenario is very highly suspicious of really being an early stage for a `DeadPrimary` scenario, and we want to find out sooner rather than later. Here, we restart all replicas.” https://github.com/vitessio/vitess/pull/18626

>  * `vtorc`: implement `restartDirectReplicas` as response to `UnreachablePrimary` [#18626](https://github.com/vitessio/vitess/pull/18626) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Deprecated VTGate metrics and engine primitive methods removed

- Affects: setups using deprecated VTGate metrics or engine primitive methods (Inferred)

The removed VTGate metrics include `QueriesProcessed`.

Evidence: “This PR removed the deprecated metrics on VTGate and removed deprecated methods on engine primitive.” https://github.com/vitessio/vitess/pull/18149

>  * Remove deprecated methods and metrics [#18149](https://github.com/vitessio/vitess/pull/18149) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Multi-statement `ExecuteFetchAsDba` requests rejected

- Affects: code or automation submitting multiple semicolon-separated SQL statements to `ExecuteFetchAsDba`
- Action required: Each statement must be sent in a separate RPC call.

Previously, requests were accepted when every statement was `CREATE TABLE` or `CREATE VIEW`; that exception no longer applies.

> #### <a id="executefetchasdba-multistatement"/>ExecuteFetchAsDba No Longer Accepts Multi-Statement SQL</a>
> The `ExecuteFetchAsDba` RPC method in TabletManager now explicitly rejects SQL queries containing multiple statements (as of [PR #18183](https://github.com/vitessio/vitess/pull/18183)).
> **Impact**: Code or automation that previously passed multiple semicolon-separated SQL statements to `ExecuteFetchAsDba` will now receive an error. Each SQL statement must be sent in a separate RPC call.
> **Migration**: Split multi-statement SQL into individual RPC calls:
> ```go
> // Before (no longer works):
> ExecuteFetchAsDba("CREATE TABLE t1 (id INT); CREATE TABLE t2 (id INT);")
> // After (required in v23+):
> ExecuteFetchAsDba("CREATE TABLE t1 (id INT);")
> ExecuteFetchAsDba("CREATE TABLE t2 (id INT);")
> ```
>  * tablet manager's `ExecuteFetchAsDba` does not allow multiple queries [#18183](https://github.com/vitessio/vitess/pull/18183) 

### Correct error codes from the `vttablet` `tabletmanager` client

- Affects: setups using the `vttablet` gRPC `tabletmanager` client
- Action required: External automation relying on google-gRPC error codes must be updated to inspect codes with `vterrors.Code(err)` and use `vtrpcpb.Code` values, such as `vtrpcpb.Code_NOT_FOUND`.

When a `tabletmanager` RPC failed to execute—for example, because `mysqld` was down—the client could report `Unknown`; now that case reports `Unavailable`, with the error message unchanged.

Evidence: “Now tablets with downed-`mysqld` will return the gRPC `Unavailable` code with the error message unchanged, for example” https://github.com/vitessio/vitess/pull/18565

> #### <a id="grpc-error-codes"/>gRPC TabletManager Error Code Changes</a>
> The `vttablet` gRPC `tabletmanager` client now returns errors wrapped by the internal `go/vt/vterrors` package ([PR #18565](https://github.com/vitessio/vitess/pull/18565)).
> **Impact**: External automation relying on google-gRPC error codes must be updated to use `vterrors.Code(err)` to inspect error codes, which returns `vtrpcpb.Code`s defined in [proto/vtrpc.proto](https://github.com/vitessio/vitess/blob/main/proto/vtrpc.proto#L60).
> **Migration**:
> ```go
> // Before:
> if status.Code(err) == codes.NotFound { ... }
> // After:
> if vterrors.Code(err) == vtrpcpb.Code_NOT_FOUND { ... }
> ```
> #### <a id="grpctmclient-err-changes"/>gRPC `tabletmanager` client error changes</a>
> The `vttablet` gRPC `tabletmanager` client now returns errors wrapped by the internal `go/vt/vterrors` package ([PR #18565](https://github.com/vitessio/vitess/pull/18565)). External automation relying on google-gRPC error codes should now use `vterrors.Code(err)` to inspect the code of an error, which returns `vtrpcpb.Code`s defined in [the `proto/vtrpc.proto` protobuf](https://github.com/vitessio/vitess/blob/main/proto/vtrpc.proto#L60).
> See [Breaking Changes](#grpc-error-codes) for migration guidance.
>  * `tabletmanager`/`grpctmclient`: use `vtrpcpb`/gRPC error codes [#18565](https://github.com/vitessio/vitess/pull/18565)

### GTID parsing method adds commit and sequence values

- Affects: code calling the GTID parsing APIs directly
- Action required: Updates to direct callers of the GTID parsing APIs to accommodate the changed method signature

Previously, `BinlogEvent.GTID()` returned only the GTID, `hasBegin`, and an error; its signature now also includes `lastCommitted` and `sequenceNumber`.

> Several GTID-related API signatures changed in [PR #18196](https://github.com/vitessio/vitess/pull/18196) as part of GTID performance optimizations:
> **Changed**: `BinlogEvent.GTID()` method signature
> **Impact**: Code directly using the GTID parsing APIs may need updates. Most users are unaffected as these are internal APIs.

### `key.GenerateShardRanges()` now accepts a hex-width parameter

- Affects: code that calls `key.GenerateShardRanges()` directly
- Action required: Direct callers must update their calls to pass the new parameter.

Previously, callers supplied only the shard count. The new `hexChars` value controls the hex width of generated shard names.

> #### <a id="generateshardranges-api"/>GenerateShardRanges API Signature Change</a>
> The `key.GenerateShardRanges()` function signature changed in [PR #18633](https://github.com/vitessio/vitess/pull/18633) to add a new `hexChars int` parameter controlling the hex width of generated shard names.
> **Impact**: Code calling `GenerateShardRanges()` directly must be updated to pass the new parameter.
> ---

### Vitess query buffering during eligible DDL changes

- Affects: eligible `INSTANT` DDL migrations and range partition rotation changes
- Action required: An instant-able migration with `--prefer-instant-ddl`, or a range partition rotation change.

This reduces query failures while the schema change runs and is transparent to applications.

> Query buffering has been implemented for `INSTANT` DDL operations ([PR #17945](https://github.com/vitessio/vitess/pull/17945)), reducing query failures during schema changes.
> - Automatic buffering of queries during INSTANT DDL execution
> - Transparent to applications
> **Impact**: Applications experience fewer query errors during schema changes, improving availability during DDL operations.
>  * Query buffering, terminating blocking transactions for `INSTANT` DDL and other "special plans" [#17945](https://github.com/vitessio/vitess/pull/17945)

### Lock-holder termination before eligible special-operation DDL

- Affects: eligible instant DDL and range partition rotation operations with nonzero `--force-cut-over-after`

Previously, these operations ran the DDL directly. Now, queries and transactions holding locks or metadata locks on the migrated table are terminated before `ALTER TABLE`; the flag's value does not change whether this happens, and termination is skipped while an atomic transaction is pending for the table.

Evidence: “In addition, if `--force-cut-over-duration` flag is nonzero, and irrespective of its value, we also issue a one time termination for queries and transactions holding locks or metadata locks on the migrated table, just before issuing the `ALTER TABLE` statement.” https://github.com/vitessio/vitess/pull/17945

> - Forced termination of blocking transactions
>  * Query buffering, terminating blocking transactions for `INSTANT` DDL and other "special plans" [#17945](https://github.com/vitessio/vitess/pull/17945)

### Cgroup CPU and memory metrics now require cgroup v2

- Affects: Linux installations using cgroup v1

Previously, Vitess collected CPU and memory metrics from both cgroup v1 and v2. With cgroup v1 or failed cgroup metric setup, it now warns and uses host metrics; cgroup v2 continues to use cgroup metrics.

Evidence: “func getCgroup1CpuUsage() (uint64, error) {” https://github.com/vitessio/vitess/pull/18801

> #### Observability
>  * [release-23.0] Improve cgroup metric management (#18791) [#18801](https://github.com/vitessio/vitess/pull/18801) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VStream diagnostic logging changes

- Affects: clusters using VStream (Inferred)

The added lines are intended to help localize frequent EOF errors in a cluster. One VStream error log changes from `Errorf` to `Infof`.

>  * VStream API: add additional log lines for debugging frequent eof errors during catchup [#18124](https://github.com/vitessio/vitess/pull/18124)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VReplication stops indefinitely retrying MySQL error 1419

- Affects: VReplication workflows

The error can occur when a workflow creates a `DEFINER` trigger; before this change, VReplication did not handle it and could retry the workflow indefinitely.

Evidence: “name: "SQL error with ERBinlogCreateRoutineNeedSuper"” https://github.com/vitessio/vitess/pull/18820

>  * [release-23.0] VReplication: Treat ER_BINLOG_CREATE_ROUTINE_NEED_SUPER as unrecoverable (#18784) [#18820](https://github.com/vitessio/vitess/pull/18820) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTTablet reports stale replica heartbeats as errors

- Affects: replica tablets using the heartbeat reader (Inferred)

When MySQL stalls and no heartbeat has been successfully read for more than two configured heartbeat intervals, `vttablet` could keep reporting MySQL as healthy. Now the heartbeat status returns an error and signals VTGate that the tablet is unhealthy.

Evidence: “The workaround I'm proposing is to change `*heartbeatReader.Status` func to just disregard any previously read heartbeat data if the last time the heartbeat was successfully read is older than two times the configured interval. It feels a bit hacky, but works reliably in identifying that MySQL is unhealthy and very quickly signals to `vtgate` that the tablet is unhealthy.” https://github.com/vitessio/vitess/pull/17883

>  * Fix `vttablet` not being marked as not serving when MySQL stalls [#17883](https://github.com/vitessio/vitess/pull/17883)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `InitAuthServerClientCert` now takes an explicit CA value

- Affects: Go code calling `go/mysql`'s public `InitAuthServerClientCert` function, including the VTGate plugin initializer (Inferred)

Previously, the function looked up `mysql-server-ssl-ca`, falling back to `mysql_server_ssl_ca` when the first flag was empty. It now takes the CA value as a second argument, which the VTGate plugin initializer supplies through `vtgate.GetMysqlServerSSLCA()`.

Evidence: “func InitAuthServerClientCert(clientcertAuthMethod string, caValue string) {” https://github.com/vitessio/vitess/pull/18508

>  * Set up better dependency checks [#18508](https://github.com/vitessio/vitess/pull/18508)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Added `lite:mysql80` Docker image build

- Affects: builds selecting the `lite:mysql80` image (Inferred)

Previously, the lite image build matrix included `latest`, `mysql84`, and `percona80`, but not `mysql80`.

>  * build `lite:mysql80` docker image [#18692](https://github.com/vitessio/vitess/pull/18692)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Forced cut-over now terminates table lock holders during `RENAME`

- Affects: Online DDL migrations using forced cut-over

Previously, a long-running query could start between the pre-cut-over termination and the `RENAME`, leaving it holding a lock on the migrated table. The additional termination pass excludes the lock and rename connections.

Evidence: “With forced cut-over, we now kill queries and transactions holding onto locks on migrated table, not only _before_ the cut-over, but also while the `RENAME` is being applied. Without this, there's a race condition where a long running query could start running just after queries are killed and right before the `RENAME` starts running.” https://github.com/vitessio/vitess/pull/18423

>  * Online DDL cutover enhancements [#18423](https://github.com/vitessio/vitess/pull/18423)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Online DDL cut-over lock-wait timeouts shortened

- Affects: Online DDL cut-over migrations (Inferred)

The locking connection's timeout changes from five times `onlineDDL.CutOverThreshold` to three times the threshold. The rename connection's timeout changes from five times the threshold times four to twice the threshold.

Evidence: “- Sanitized and reduced query and lock timeouts during the cut-over. There were some excessive timeouts, notably an overlooked `5*onlineDDL.CutOverThreshold*4` value which evaluates to `5min` on a `15s` timeout.” https://github.com/vitessio/vitess/pull/18423

>  * Online DDL cutover enhancements [#18423](https://github.com/vitessio/vitess/pull/18423)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTOrc `database_instance` location fields replaced by informational `cell`

- Affects: VTOrc installations using the `database_instance` table (Inferred)

The removed fields were not used in logic; `physical_environment` and `region` were always undefined or empty, while `data_center` matched the tablet cell. The new `cell` field is not used in logic yet.

Evidence: “The 3 x fields are replaced by the new `cell` field. It isn't used in any logic yet - for now it is informational” https://github.com/vitessio/vitess/pull/18339

>  * `vtorc`: cleanup `database_instance` location fields [#18339](https://github.com/vitessio/vitess/pull/18339)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTOrc `database_instance` schema replaces location fields with `cell`

- Affects: VTOrc installations using the `database_instance` table (Inferred)

Previously, the table had `data_center`, `physical_environment`, and `region` fields, which the pull request says were not used in logic. The replacement `cell` field is informational and is not used in logic yet.

Evidence: “The 3 x fields are replaced by the new `cell` field. It isn't used in any logic yet - for now it is informational” https://github.com/vitessio/vitess/pull/18339

>  * `vtorc`: cleanup `database_instance` location fields [#18339](https://github.com/vitessio/vitess/pull/18339)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTOrc's aggregated discovery metrics API removed

- Affects: Integrations using VTOrc's `/api/aggregated-discovery-metrics` endpoint (Inferred)

The endpoint previously returned JSON of aggregated discovery samples over a 120-second period.

Evidence: “This PR drops the undocumented VTOrc API endpoint `/api/aggregated-discovery-metrics`” https://github.com/vitessio/vitess/pull/18672

>  * `vtorc`: remove aggregated discovery metrics HTTP API, cleanup code [#18672](https://github.com/vitessio/vitess/pull/18672)
> #### <a id="aggregated-discovery-metrics-api-removed"/>Aggregated Discovery Metrics HTTP API removed</a>
> VTOrc's undocumented `/api/aggregated-discovery-metrics` HTTP API endpoint was removed ([PR #18672](https://github.com/vitessio/vitess/pull/18672)). The list of documented VTOrc APIs can be found [here](https://vitess.io/docs/current/reference/vtorc/ui_api_metrics/#apis).
> We recommend using the standard VTOrc metrics to gather the same metrics. If you find that a metric is missing in standard metrics, please open an issue or PR to address this.

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

## Deprecations (15)

### Underscore-form CLI flags deprecated

- Affects: underscore-form names of the 989 migrated CLI flags
- Removed in: v25.0.0

Previously, a flag spelling that did not match the form bound to Viper could break or do nothing. In v23.0.0 and v24.0.0, both forms are accepted and normalized to dashed names; underscore forms are hidden from help and produce a deprecation message.

> ### <a id="flag-naming-convention-migration"/>Flag Naming Convention Migration</a>
> Vitess v23.0.0 includes a major standardization of CLI flag naming conventions across all binaries. **989 flags** have been migrated from underscore notation (`flag_name`) to dash notation (`flag-name`) in [PR #18280](https://github.com/vitessio/vitess/pull/18280) and related PRs.
> #### Backward Compatibility
> - **v23.0.0 and v24.0.0**: Both underscore and dash formats are supported. Underscore format is **deprecated** but functional.
> - **v25.0.0**: Underscore format will be **removed**. Only dash format will be accepted.
> #### Automatic Normalization
> Flag normalization happens automatically at the `pflag` level ([PR #18642](https://github.com/vitessio/vitess/pull/18642)), so both formats are accepted without requiring code changes in v23/v24.
> #### Example Flag Renames
> Common flags affected (full list of 989 flags available in PR #18280):
> **Backup flags**:
> - `--azblob_backup_account_name` → `--azblob-backup-account-name`
> - `--s3_backup_storage_bucket` → `--s3-backup-storage-bucket`
> - `--xtrabackup_root_path` → `--xtrabackup-root-path`
> **Replication flags**:
> - `--heartbeat_enable` → `--heartbeat-enable`
> - `--replication_connect_retry` → `--replication-connect-retry`
> **gRPC flags** ([PR #18009](https://github.com/vitessio/vitess/pull/18009)):
> - All gRPC-related flags standardized (30+ flags)
> #### Action Required
> Users should update configuration files, scripts, and automation to use dash-based flag names before upgrading to v25.0.0. The migration is backward compatible in v23 and v24, allowing gradual updates.

### VTOrc adds the `DiscoveryInstanceTimings` metric

- Affects: VTOrc deployments consuming instance-discovery timing metrics (Inferred)
- Removed in: v24+

Both metrics receive instance-discovery action timings in this release. The old `DiscoverInstanceTimings` metric is planned for deprecation in v24 and deletion in v24+.

> | `vtorc`   | `DiscoverInstanceTimings` | Replaced by `DiscoveryInstanceTimings` | [#18406](https://github.com/vitessio/vitess/pull/18406) |
>  * `vtorc`: update metrics in e2e test `TestAPIEndpoints` [#18406](https://github.com/vitessio/vitess/pull/18406)

### `/api/replication-analysis` deprecated in favor of `/api/detection-analysis`

- Affects: VTOrc API clients using either analysis endpoint
- Removed in: v24+

Both endpoints use the same handler in this release and return the same response format. The old endpoint remains available until its planned removal in v24+.

> #### <a id="replication-analysis-api-deprecation"/>`/api/replication-analysis` HTTP API deprecation</a>
> The `/api/replication-analysis` HTTP API endpoint is now deprecated and is replaced with `/api/detection-analysis` ([PR #18615](https://github.com/vitessio/vitess/pull/18615)), which currently returns the same response format.
> **Timeline**: The `/api/replication-analysis` endpoint will be removed in a future version. Users should migrate to `/api/detection-analysis`.

### Managed MySQL authentication defaults to `caching_sha2_password`

- Affects: Managed MySQL instances on MySQL 8.0.26 or later
- Removed in: not stated

Previously, the default was `mysql_native_password`, which remains enabled for backwards compatibility. The note recommends `caching_sha2_password` for replication users configured with an explicit password after upgrading to v23.

> #### <a id="mysql-caching-sha2-password"/>Managed MySQL configuration defaults to caching-sha2-password</a>
> The default authentication plugin for MySQL 8.0.26 and later is now `caching_sha2_password` instead of `mysql_native_password` ([PR #18010](https://github.com/vitessio/vitess/pull/18010)). This change is made because `mysql_native_password` is deprecated and removed in future MySQL versions. `mysql_native_password` is still enabled for backwards compatibility.
> This change specifically affects the replication user. If you have a user configured with an explicit password, it is recommended to make sure to upgrade this user after upgrading to v23 with a statement like the following:
> ```sql
> ALTER USER 'vt_repl'@'%' IDENTIFIED WITH caching_sha2_password BY 'your-existing-password';
> ```
> In future Vitess versions, the `mysql_native_password` authentication plugin will be disabled for managed MySQL instances.
>  * Switch to using the default caching sha2 password [#18010](https://github.com/vitessio/vitess/pull/18010)

### Migrated gRPC flags use dashed names

- Affects: setups using gRPC flags such as `--grpc-auth-mode` and `--grpc-bind-address`
- Removed in: v25

These flags previously used underscores; their dashed names are now available, while the underscore variants remain supported but are deprecated and hidden.

>  * Refactor grpc flags for standardized naming  [#18009](https://github.com/vitessio/vitess/pull/18009)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Underscore flag names deprecated

- Affects: setups using the migrated logging, database, tablet-manager gRPC, or statistics flags
- Removed in: v25

These flags used underscore names before the refactor. Dashed names are available now; underscore variants still work but are deprecated and hidden.

>  * Refactor flags - Part 3 [#18064](https://github.com/vitessio/vitess/pull/18064)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Commonly reused Vitess flags migrate to dashed names

- Affects: setups using migrated backup-storage, MySQL, topology, or tablet flags
- Removed in: v25

These flags previously used underscore names. Their dashed names are now available, while the underscore variants remain supported but are deprecated and hidden.

>  * Refactor flags - Part 4 [#18095](https://github.com/vitessio/vitess/pull/18095)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Underscore CLI flag forms deprecated

- Affects: CLI flags changed by the flag refactor
- Removed in: v25

In v23, both dashed and underscore forms are accepted.

>  * Merge flags-refactor branch to main [#18280](https://github.com/vitessio/vitess/pull/18280)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Dashed names added for `vtctld` flags

- Affects: setups using `vtctld` flags such as `--disable-active-reparents` and `--opentsdb-uri`
- Removed in: not stated

The flags previously used underscore naming. Their dashed forms are now available; the underscore forms remain supported but are deprecated and hidden.

>  * Refactor `vtctld` flags to follow standard naming convention [#18294](https://github.com/vitessio/vitess/pull/18294)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vitess flags migrated to dashed names

- Affects: setups using the migrated DBA pool, VTGate/Vtctld gRPC, client protocol, query logging, external-topology, or Ceph backup-storage flags
- Removed in: not stated

The underscore-named variants remain supported, but are now deprecated and hidden.

>  * Flags Refactor - Part 5 [#18296](https://github.com/vitessio/vitess/pull/18296)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Migrated `vtcombo` flags use dashed names

- Affects: setups using the migrated flags in the `vtcombo` binary
- Removed in: not stated

Their underscore-named forms remain supported, but are deprecated and hidden.

>  * Refactor `vtcombo` flags - Part 2 [#18297](https://github.com/vitessio/vitess/pull/18297)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `vtcombo` flags move to dashed names

- Affects: migrated flags in the `vtcombo` binary
- Removed in: not stated

Before the refactor, these flags used underscores. Their dashed names are now available; underscore variants remain supported but are deprecated and hidden.

>  * Refactor `vtcombo` flags - Part 3 [#18298](https://github.com/vitessio/vitess/pull/18298)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Affected command-line flags standardized to dashed names

- Affects: Vitess deployments invoking `mysqlctld`, `mysqlctl`, or `vtbackup` commands with the affected flags
- Removed in: v25

The corresponding flags previously used underscore spellings, including `init_db_sql_file`, `socket_file`, and `wait_time`.

>  * Refactor `mysqlctld` flags to follow standardized naming [#18432](https://github.com/vitessio/vitess/pull/18432)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### 47 `vtcombo` flags gain dashed names

- Affects: `vtcombo` configurations using the migrated flags
- Removed in: v25

The affected flags previously used underscores; the PR lists `--relay-log-max-items` and `--restore-from-backup` among the dashed spellings.

>  * Refactor `vtcombo` flags - Part 4 [#18422](https://github.com/vitessio/vitess/pull/18422) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `/api/replication-analysis` deprecated in favor of `/api/detection-analysis`

- Affects: VTOrc clients using `/api/replication-analysis` (Inferred)
- Removed in: not stated

The replacement endpoint returns the same response format.

>  * `vtorc`: rename `ReplicationAnalysis` -> `DetectionAnalysis` [#18615](https://github.com/vitessio/vitess/pull/18615)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

## Important fixes (49)

### mysqlctl environment variables propagated to mysqld

- Affects: mysqlctl deployments with environment variables such as a non-UTC `TZ` setting

Previously, `TZ` set for mysqlctl was ignored by mysqld, so timezone settings from the environment did not affect `DATETIME` interpretation. mysqld now inherits those variables, and deployments relying on the old behavior may see `DATETIME` values interpreted differently.

> Fixed a bug where environment variables like `TZ` were not propagated from mysqlctl to the mysqld process ([PR #18561](https://github.com/vitessio/vitess/pull/18561)).
> As a result, timezone settings from the environment were previously ignored. Now mysqld correctly inherits environment variables.
> ⚠️ **Deployment Impact**: Deployments that relied on the old behavior and explicitly set a non-UTC timezone may see changes in how DATETIME values are interpreted. To preserve compatibility, set `TZ=UTC` explicitly in MySQL pods.
>  * Pass envs to mysqld process [#18561](https://github.com/vitessio/vitess/pull/18561) 

### Underscored CLI flag handling fixed

- Affects: CLI flags spelled with underscores, especially those bound to Viper

When an underscored flag was passed, Viper could miss its value because it was bound to a different flag object, so the setting could break or have no effect. Underscored names are now normalized to dashed names during parsing, so Viper receives the supplied value; they still produce a deprecation message.

>  * Normalize flag names at the `pflag` level. [#18642](https://github.com/vitessio/vitess/pull/18642) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Evalengine tuple simplification preserves multi-column results

- Affects: evalengine tuple simplification for multi-column results, including CASE expressions in join conditions (Inferred)

Before this fix, the simplifier wrapped multi-column results in a single `Literal`; a CASE expression in a join condition is a demonstrated scenario.

>  * fix: Preserve multi-column TupleExpr in tuple simplifier [#18216](https://github.com/vitessio/vitess/pull/18216)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Evalengine comparison crashes fixed

- Affects: evalengine comparisons involving NULLs or JSON, ENUM, SET, or geometry values (Inferred)

Comparisons involving NULLs or types evalengine did not handle properly, including JSON, ENUM, and SET, could cause a panic. The affected NULL checks and type-handling cases are fixed, so the described unexpected-type crashes no longer occur.

>  * Fix evalengine crashes on unexpected types [#18254](https://github.com/vitessio/vitess/pull/18254) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### vtgate base-table insertion fixed for ignored vindex rows

- Affects: `INSERT IGNORE` statements using vindexes

Previously, when a vindex reported an ignored row during `INSERT IGNORE`, a NULL vindex column could keep the row out of the lookup table and cause the base-table insertion to be skipped. Now vtgate can insert the row into the base table despite the vindex reporting it ignored.

>  * bugfix: INSERT IGNORE not inserting rows [#18151](https://github.com/vitessio/vitess/pull/18151)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### LIMIT/OFFSET preserved when merging UNION queries

- Affects: UNION queries Vitess merges that use LIMIT or OFFSET (Inferred)

Before this fix, merged UNION queries could miss their LIMIT/OFFSET.

>  * fix: keep LIMIT/OFFSET even when merging UNION queries [#18361](https://github.com/vitessio/vitess/pull/18361)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Empty-input scalar aggregation literal results fixed

- Affects: queries using scalar aggregations with literal values (Inferred)

When the input was empty—for example, because a subquery's `WHERE` condition returned no rows—the literal was treated like a column reference and the example returned (`NULL`, `0`) instead of (`1`, `0`). The literal is now evaluated, producing the MySQL-compatible result (`1`, `0`).

>  * Fix scalar aggregation with literals in empty result sets [#18477](https://github.com/vitessio/vitess/pull/18477)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Prepared-statement field list fix for simple projections

- Affects: prepared statements with simple projections using the names-only optimization (Inferred)

When `GetFields` called `buildFields` directly for a prepared statement with a simple projection, the names-only case produced an empty fields list.

>  * Fix for simple projection showing no fields [#18489](https://github.com/vitessio/vitess/pull/18489)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vitess planner fixes column counts for derived-table `GROUP BY` plans

- Affects: queries with `GROUP BY` only on top of derived tables

For these queries, the planner could report an incorrect projection width, allowing extraneous columns to leak into the result.

>  * bugfix: Plan group by only on top of derived tables correctly [#18505](https://github.com/vitessio/vitess/pull/18505)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### [Silent failure] Throttler metric aggregation recognizes gRPC dial errors

- Affects: setups using throttler metric aggregation

When a metric probe’s gRPC connection failed with `Unavailable` or `DeadlineExceeded`, aggregation treated it as an unhealthy state, causing throttler users such as `VReplication` to throttle.

>  * Properly handle grpc dial errors in the throttler metric aggregation [#18073](https://github.com/vitessio/vitess/pull/18073)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Tablet throttler keeps its `SrvKeyspace` watch after errors

- Affects: setups using the tablet throttler's `SrvKeyspace` watch

When the watch callback received an error, its listener was removed, so the throttler stopped reading `SevKeyspace` changes until it called `Open()` again, normally when the tablet type changed. The watch now stays active after an error.

>  * Throttler: keep watching topo even on error [#18223](https://github.com/vitessio/vitess/pull/18223) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Atomic Copy post-copy action execution fixed

- Affects: workflows using Atomic Copy with post-copy actions for copied tables

`copyAll()` could miss executing post-copy actions for a table. Executed action records are now removed from `_vt.post_copy_action` along with the corresponding `_vt.copy_state` record.

>  * Perform post copy actions in atomic copy [#18411](https://github.com/vitessio/vitess/pull/18411)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Replication tracker resets the replication-lag metric for primary tablets

- Affects: VTTablet setups using the replication tracker (Inferred)

A previous lag value could remain after promotion to primary or when `Status()` ran while the tracker was primary or in `Disable` mode. The metric now resets to zero in those cases.

>  * [release-23.0] repltracker: reset replica lag when we are primary (#18800) [#18807](https://github.com/vitessio/vitess/pull/18807)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### [Silent failure] Query consolidator waiter-cap fallback fixed

- Affects: query-serving setups using query consolidation (Inferred)

When the waiter cap was exceeded, the consolidator could return an empty result without an error. Now the query runs independently and returns its result.

>  * [release-23.0] Fix bug where query consolidator returns empty result without error when the waiter cap exceeded (#18782) [#18833](https://github.com/vitessio/vitess/pull/18833) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### [Silent failure] Incomplete range coverage in `GenerateShardRanges` fixed

- Affects: shard-name generation with `GenerateShardRanges` (Inferred)

For some shard counts, such as 49, the generated shard names did not cover the full range. They now cover the full keyspace.

>  * fix: Fix `GenerateShardRanges` returning shard names that don't cover the full range [#18641](https://github.com/vitessio/vitess/pull/18641) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Online DDL progress corrected to 100% at `ready_to_complete`

- Affects: Online DDL migrations that reach `ready_to_complete`

At `ready_to_complete`, table-row estimation could leave reported progress at 87% or another value far from 100%.

>  * Online DDL: set migration's progress to 100% once it's `ready_to_complete` [#18544](https://github.com/vitessio/vitess/pull/18544)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Impossible-query formatting fixed for `UNION` queries

- Affects: impossible-query column metadata fetches for `UNION` statements

When impossible queries fetched column names and types for a `UNION` without fetching data, the formatter did not reliably add the no-rows condition. It now includes `where 1 != 1` in the formatted query, including its union branches.

>  * bugfix: Fix impossible query for UNION [#18463](https://github.com/vitessio/vitess/pull/18463)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VStreamer filters send both images for matching non-sharded updates

- Affects: Non-sharded VStreamer workflows other than `MoveTables` and `Reshard`, with filters on row columns

When an update changed columns used by the filter and either the `Before` or `After` image matched, only the passing image was sent, so downstream consumers could interpret the update as an insert or delete.

>  * VStreamer: change in filter logic [#18319](https://github.com/vitessio/vitess/pull/18319)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTOrc `SkippedRecoveries` metric gains a reason dimension

- Affects: VTOrc installations that emit or consume `SkippedRecoveries` metrics

When VTOrc skips a recovery, the metric identifies reasons such as `ERSDisabled`, `ReplicationLagHigh`, and `NotEnoughReplicas`, and tracks previously missing skipped-recovery cases.

> |          Name                         |   Dimensions                                  |                   Description                        |                           PR                            |
> |:-------------------------------------:|:---------------------------------------------:|:----------------------------------------------------:|:-------------------------------------------------------:|
> | `SkippedRecoveries`                   | `RecoveryName`, `Keyspace`, `Shard`, `Reason` | Count of skipped recoveries with reason tracking.    | [#18644](https://github.com/vitessio/vitess/pull/18644) |
> **`SkippedRecoveries` Reasons**: The `Reason` dimension tracks why recoveries were skipped (e.g., "ERSDisabled", "ReplicationLagHigh", "NotEnoughReplicas"), providing actionable insights for operational troubleshooting.
>  * `vtorc`: add `Reason` and missing cases to `SkippedRecoveries` metric [#18644](https://github.com/vitessio/vitess/pull/18644)

### `CREATE PROCEDURE` normalization fix for body `SET` statements

- Affects: stored procedures containing `SET` statements (Inferred)

During planner normalization, `SET y = 4` in a procedure body was changed to `SET @@y = 4`. Vitess now skips that normalization for `CREATE PROCEDURE` statements, leaving the body’s `SET` statement unchanged.

> - Support for `SET` statements within procedure bodies
>  * Fix `SET` and `START TRANSACTION` in create procedure statements [#18279](https://github.com/vitessio/vitess/pull/18279)

### Vitess fixes `MEMBER OF` precedence with `AND`

- Affects: JSON queries using `MEMBER OF`

When these queries combined `MEMBER OF` with `AND` or other boolean logic, incorrect precedence could cause Vitess to send incorrect queries to the underlying MySQL instances. They are now evaluated using the corrected precedence.

> - Fixed `MEMBER OF` operator precedence with `AND` ([PR #18237](https://github.com/vitessio/vitess/pull/18237))
> - Ensures correct query evaluation when combining JSON operations with boolean logic
>  * make sure to give MEMBER OF the correct precedence [#18237](https://github.com/vitessio/vitess/pull/18237)

### Orphaned serving records cleared for deleted keyspaces

- Affects: keyspace rebuilding with `RebuildKeyspace` (Inferred)

Previously, residual serving records left after deleting a keyspace caused `vtgate` to crash when `RebuildKeyspace` encountered the deleted keyspace. It now clears those records instead.

>  * Add code to clear orphaned files for a deleted keyspace [#18098](https://github.com/vitessio/vitess/pull/18098) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vreplication restart on failed Online DDL table rename

- Affects: Online DDL cut-overs that stop vreplication before renaming tables (Inferred)

If the remaining `RENAME` fails after vreplication has been stopped, the cut-over attempts to restart vreplication for that workflow.

> #### Online DDL
>  * Online DDL: resume vreplication after cut-over/RENAME failure [#18428](https://github.com/vitessio/vitess/pull/18428) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### vtgate plan keys now distinguish tablet types

- Affects: vtgate workloads that plan and cache queries for different tablet types, such as primary and replica (Inferred)

When vtgate reused a plan whose key did not distinguish tablet types, it could reuse an incorrect plan.

>  * Fix: Add tablet type to the plan key [#18155](https://github.com/vitessio/vitess/pull/18155)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Repeated refresh notifications from resilient topology watchers fixed

- Affects: setups using resilient topology watchers (Inferred)

During topology outages with cached data available, the watcher notified listeners on every refresh even while serving cached data. Now listeners remain silent, and `Get` operations continue working from the cache.

>  * Fix watcher storm during topo outages [#18434](https://github.com/vitessio/vitess/pull/18434)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Traffic mirroring support for `dual` tables

- Affects: setups using traffic mirroring for queries involving `dual` tables (Inferred)

Previously, a mirrored query involving a `dual` table with no mirror-rule entry could panic. The query is now processed without mirroring the `dual` table.

>  * go/vt/vtgate: handle dual tables in traffic mirroring [#18458](https://github.com/vitessio/vitess/pull/18458)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vitess fixes column alias handling in `UNION` planning

- Affects: queries using column aliases in `UNION`s (Inferred)

When planning a `UNION`, Vitess had difficulty handling column aliases; the example uses aliases in both branches of a `UNION ALL`.

>  * fix: handle aliases in UNIONs better [#18484](https://github.com/vitessio/vitess/pull/18484)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vitess routed-table foreign key handling fixed

- Affects: setups with foreign-key relations and routed tables

After MoveTables with SwitchWrites, a child table in the source keyspace could find its routed parent in the target keyspace, causing the parent to include a source-keyspace child and creating broken foreign key references. Now, foreign key relationships are added only for routed tables in the target keyspace, not for source-keyspace tables.

>  *  Fix foreign key relation with routed tables [#18537](https://github.com/vitessio/vitess/pull/18537)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Invalid `srvtopo` cache entries now refresh

- Affects: `srvtopo` lookups

Within the cache interval, a cached error could be returned immediately instead of querying the source again, causing calls such as `show vitess_shards` to fail with context cancellation or timeouts. Now an invalid cache entry or cached error triggers a refresh and waits for fresh results; a still-valid cached value is returned without another query.

>  * `srvtopo/query.go`: Do not query errors, fall back and refresh [#18589](https://github.com/vitessio/vitess/pull/18589)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vitess adds aliases to self-referential foreign-key verification queries

- Affects: self-referential foreign keys using the `RESTRICT` action

When Vitess plans verification queries for these foreign keys, the queries self-joined the table without an alias for the joined table and failed on MySQL. They now include aliases for joined tables, avoiding that execution failure.

>  * Introduce aliases for foreign keys verify operations [#18601](https://github.com/vitessio/vitess/pull/18601)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Connection-pool wait hang during close fixed

- Affects: setups using the connection pool (Inferred)

When a connection is requested as the pool closes, the wait could hang and prevent normal operations such as PRS/ERS from succeeding.

>  * CONNPOOL: Fix race condition when waiting for connection [#18713](https://github.com/vitessio/vitess/pull/18713)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VDiff extra-row reconciliation panic fixed

- Affects: VDiff runs that reconcile extra rows

VDiff could panic when extra source or target rows exceeded `--max-extra-rows-to-compare`, or when the source and target returned different numbers of rows. Reconciliation now uses the lengths of the extra-row difference slices, avoiding the out-of-bounds access in these cases.

> #### VDiff
>  * Fix a panic in VDiff when reconciling extra rows. [#18585](https://github.com/vitessio/vitess/pull/18585) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VStream catchup early EOF fixed

- Affects: VStream API catchup

With multiple tables, catchup could return `EOF` before processing updates or deletes for tables or rows already copied during the copy phase. Resetting the fastforward-set `stopPos` prevents this early exit so catchup can process those transactions.

>  * VStream API: Reset stopPos in catchup [#18119](https://github.com/vitessio/vitess/pull/18119)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `vcopier` closed-channel panic no longer crashes tablets

- Affects: tablet workloads using `vcopier` copy tasks

If a `copyAll` callback sent a result to `resultCh` after it was closed—possibly when the context exceeded its deadline—the panic could crash the tablet. The panic is now recovered and logged instead.

>  * Recover panic in vcopier due to closed channel [#18228](https://github.com/vitessio/vitess/pull/18228)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Atomic Copy table-streaming errors now propagated

- Affects: Workflows using Atomic Copy while streaming tables

When a stream error such as a connection timeout was ignored, the workflow could restart, re-enter the copy phase, and encounter duplicate-key errors. Now the error is propagated, and logs identify the failing table and configured net read and write timeouts.

>  * Atomic Copy: Handle error that was ignored while streaming tables and log it [#18313](https://github.com/vitessio/vitess/pull/18313)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Atomic Copy workflows no longer terminate on replication errors

- Affects: VReplication workflows in Atomic Copy sub-mode that have completed the copy phase and entered the `running` state

After the copy phase, a replication-phase error could terminate the workflow. Such errors are now handled normally instead.

>  * Avoid terminating atomic copy workflows on error if they are out of copy phase [#18475](https://github.com/vitessio/vitess/pull/18475)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### MariaDB rejection of reverse-workflow VReplication reads fixed

- Affects: VReplication imports from MariaDB using a reverse workflow

Reading VReplication workflows in these imports could fail because MariaDB rejected the query under stricter SQL grouping rules. The query no longer includes the `GROUP BY` clause that triggered the rejection.

>  * VReplication: Fix bug while reading _vt.vreplication record [#18478](https://github.com/vitessio/vitess/pull/18478)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VStreamer concurrent-map crash fixed

- Affects: VStreamer event parsing in `vttablet` (Inferred)

Concurrent map access during in-place GTID computation could crash `vttablet` with `fatal error: concurrent map read and map write`. VStreamer now uses `replication.AppendGTID` instead of the in-place operation.

>  * Vstreamer: idempotent gtid set (bugfix) [#18611](https://github.com/vitessio/vitess/pull/18611)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VStream retries after purged-binlog errors

- Affects: VStream requests

When a tablet lacks the requested GTID because the source purged the required binary logs, VStream previously did not ignore that tablet and therefore did not try another on that basis. It now retries with a different tablet, which may still have the requested binlog history.

>  * VStream: Try new tablet on purged binlog error [#18710](https://github.com/vitessio/vitess/pull/18710)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VReplication handling of dashed identifiers fixed

- Affects: VReplication-related operations using keyspace, database, table, or sequence names that contain dashes

These operations now escape identifiers such as the table keyspace, table name, database, and sequence before using them in SQL.

>  * [release-23.0] VReplication: Ensure proper handling of keyspace/database names with dashes (#18762) [#18773](https://github.com/vitessio/vitess/pull/18773)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### etcd2topo shutdown panic fixed

- Affects: setups using the etcd2topo server implementation

During VTGate shutdown, concurrent coroutines could access `s.cli` after `Close()` set it to `nil`, causing a nil pointer dereference panic. Such operations now return a shutdown signal (`context.Canceled`, converted to `topo.Interrupted`) instead of panicking.

>  * Make sure to check if the server is closed in etcd2topo [#18352](https://github.com/vitessio/vitess/pull/18352)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Semi-sync monitor timer shutdown deadlock fixed

- Affects: setups running the VTTablet semi-sync monitor (Inferred)

Closing the timer while holding `mu` could deadlock when `checkAndFixSemiSyncBlocked` also acquired it. The monitor now serializes timer `Open` and `Close` with a separate timer mutex and closes its ticks before acquiring `mu`.

>  * Fix deadlock in semi-sync monitor [#18276](https://github.com/vitessio/vitess/pull/18276)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTTablet semi-sync monitor deadlock fixed

- Affects: setups running the VTTablet semi-sync monitor (Inferred)

When semi-sync was blocked and the monitor closed after a heartbeat write started but before `setIsBlocked` changed monitor state, `appPool` shutdown could deadlock: it waited for the write's connection to be recycled while the write thread waited for the mutex. The monitor now recycles the connection before waiting for the mutex, so shutdown no longer waits on a connection held by a blocked write.

>  * Fix: Deadlock in `Close` and `write` in semi-sync monitor. [#18359](https://github.com/vitessio/vitess/pull/18359)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Recurring `VStreamRows` heartbeats during copy phases

- Affects: long-running `VStreamRows` operations

Previously, the heartbeat mechanism sent one heartbeat and then exited during long-running table copies, including OnlineDDL copy phases and table copying during MoveTables/Reshard.

>  * [Bugfix] Broken Heartbeat system in Row Streamer [#18390](https://github.com/vitessio/vitess/pull/18390)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### `ValidateKeyspace` deadlock fixed

- Affects: setups using `ValidateKeyspace` through `vtctldclient` (Inferred)

Result writers could block indefinitely if they started writing before a reader was active. Now, results are processed and consumed before writing begins, so writers have an active reader.

> #### vtctldclient
>  * Fix deadlock in `ValidateKeyspace` [#18114](https://github.com/vitessio/vitess/pull/18114)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### vttablet external MySQL connectivity check passes SSL configuration

- Affects: vttablet connections to external MySQL instances that require secure connections

When vttablet checked connectivity to an external MySQL instance requiring secure connections, the check failed to pass SSL configuration, and MySQL's error prevented vttablet from starting. The check now passes SSL configuration to the connection logic, avoiding that startup error.

>  * Pass SSL config when checking connection to external mysql instance [#18481](https://github.com/vitessio/vitess/pull/18481) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VTOrc prioritizes primary recovery over secondary mitigations

- Affects: VTOrc setups handling non-cluster-wide recovery analyses

When a non-cluster-wide issue such as `ReplicationStopped` was caused by a dead primary, VTOrc could lock the shard to mitigate the secondary issue before recognizing the primary failure, delaying primary recovery. VTOrc now refreshes primary information before proceeding and aborts the stale analysis if primary mitigation is needed.

>  * Vtorc: Recheck primary health before attempting a failure mitigation [#18234](https://github.com/vitessio/vitess/pull/18234)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### Vitess Java JDBC driver isolation-variable lookup fixed

- Affects: Vitess Java JDBC driver with MySQL 8.0 or MySQL 5.7.20 and later

On MySQL 8.0, initialization previously queried `tx_isolation`, which was deprecated or deleted. The replacement, `transaction_isolation`, is backwards compatible with MySQL 5.7.20 and later.

>  * Use transaction_isolation vs tx_isolation [#17845](https://github.com/vitessio/vitess/pull/17845)

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

### VDiff sidecar schema updated to improve query performance

- Affects: setups running VDiff with its internal sidecar database (Inferred)

When VDiff ran, poorly performing queries against the internal sidecar database could cause serious issues, including crashes of primary vttablets. The schema now uses `bigint(20)` for `vdiff_log.id`, `vdiff_log.vdiff_id`, and `vdiff_table.vdiff_id`, and adds the `vdiff_id_idx` index to improve some VDiff queries.

>  * Fix: Improve VDiff internal query performance [#18579](https://github.com/vitessio/vitess/pull/18579) 

https://github.com/vitessio/vitess/blob/main/changelog/23.0/23.0.0/changelog.md

## Enhancements (47)

- `ExecuteMulti` and `StreamExecuteMulti` APIs for VTGate (To use: Enable `--mysql-server-multi-query-protocol` on VTGate to use the feature.)
- Per-session transaction timeout control
- Experimental incoming query throttling framework
- Multiple lookup vindexes in one `vtctldclient` workflow (To use: Use `--params-file` with a JSON file containing lookup vindex parameters; it is mutually exclusive with `--type`, `--table-owner`, and `--table-owner-columns`.)
- Adding reference tables to existing materialize workflows (To use: `Materialize ... update` with `--add-reference-tables` and a target keyspace)
- Recursive CTE support in Vitess
- Vitess SQL parser support for `CREATE TABLE ... SELECT`
- VTGate adds a transaction counter by shard and type
- VTGate adds a deferred-optimization query counter
- Online DDL stale-migration age gauge added
- VTOrc adds the `EmergencyReparentShardDisabled` gauge
- Any-condition query logging in vtgate, vttablet, and vtcombo (To use: `--querylog-emit-on-any-condition-met` enabled (false by default))
- VTOrc startup recovery control (To use: `--allow-recovery=false`)
- Configurable hex width for `GenerateShardRanges` shard labels (To use: `--chars` flag)
- Support for `ALTER VITESS_MIGRATION ... POSTPONE COMPLETE` commands
- Parser support for singular `VALUE` in `INSERT` and `REPLACE`
- Per-keyspace and per-shard control for VTOrc emergency reparenting (To use: the `vtctldclient SetVtorcEmergencyReparent` command with `--keyspace`, optionally `--shard`, and `--enabled=true` or `--enabled=false`)
- VTGate support for tuple bind variables in `IN` filters
- Resharding journal `val` column widened to `longblob`
- Empty-password credentials accepted for unmanaged tablets
- Configurable tablet draining for the `xtrabackup` backup engine (To use: `--xtrabackup-should-drain` set to `true`)
- Parsed MySQL connection attributes retained for server use
- MySQL client support for handshake connection attributes (To use: Use `ConnectWithAttributes` with a `ConnectionAttributes` map.)
- Custom logger support for discovery (To use: `WithLogger` option)
- VTTablet logs primary-serving binlog watcher closures
- gRPC servers now report ORCA backend metrics (To use: Enable `--grpc-enable-orca-metrics` (default: false).)
- Query logging gains an execution-time threshold (To use: A `--querylog-time-threshold` duration in `time.ParseDuration` format; it is not useful for streaming queries.)
- Online DDL readiness-change logging
- Example Vitess Operator CRD schema adds `fetchCredentials`
- Query planner optimization-reason debug logs (To use: `DebugOperatorTree = true`)
- SQL parser support for `SET NAMES binary`
- Incoming query throttling for tablets (To use: Throttler enabled flag in `/config/throttler-config.json`)
- Query throttler dry-run mode
- New TMC RPCs batch sequence handling in `switchwrites`
- VStream table names without the keyspace (To use: Set the `ExcludeKeyspaceFromTableName` VStream flag)
- VStreamer uses schema for compatible DDL changes
- Session closing exposed through `vtgateconn` gRPC
- Configurable gRPC fail-fast for vtgate connections (To use: Set `vtgate-grpc-fail-fast` to `true` to enable gRPC fail-fast; it defaults to `false`.)
- vtgateconn gRPC flag registration exposed to external callers
- Single-call MySQL replication restart (To use: A client that calls the added `RestartReplication` method)
- `SchemaDiff` adds a foreign-key impact listing
- Per-shard completion for Online DDL migrations
- Selective table copying in VStream (To use: Set `TablesToCopy` in `VStreamOptions` or `VStreamFlag`)
- `vstreamer` emits `PREVIOUS_GTIDS` VEvents
- GTID metadata added to `vstreamer` events
- `operator.yaml` schema support for `mysqldExporter.extraFlags` (To use: `mysqldExporter.extraFlags` in `operator.yaml`, defined as an object with string-valued additional properties)
- `mysqldExporter` no longer requires `resources`

## Dependency updates (17)

17 more, none mentioning security

- `* Build mysql80 Dockerfile with go1.25.1 [#18694](https://github.com/vitessio/vitess/pull/18694)`
- `` * [main] Upgrade the Golang version to `go1.24.2` [#18093](https://github.com/vitessio/vitess/pull/18093) ``
- `` * [main] Upgrade the Golang version to `go1.24.3` [#18241](https://github.com/vitessio/vitess/pull/18241) ``
- `` * [main] Upgrade the Golang version to `go1.24.4` [#18327](https://github.com/vitessio/vitess/pull/18327) ``
- `` * [main] Upgrade the Golang version to `go1.25.0` [#18573](https://github.com/vitessio/vitess/pull/18573) ``
- `` * [main] Upgrade the Golang version to `go1.25.1` [#18622](https://github.com/vitessio/vitess/pull/18622) ``
- `* Bump subset of Go dependencies [#18577](https://github.com/vitessio/vitess/pull/18577)`
- `* Upgrade the Golang Dependencies [#18082](https://github.com/vitessio/vitess/pull/18082)`
- `* Bump golang.org/x/net from 0.37.0 to 0.38.0 [#18177](https://github.com/vitessio/vitess/pull/18177)`
- `* Upgrade the Golang Dependencies [#18266](https://github.com/vitessio/vitess/pull/18266)`
- `* Bump github.com/go-viper/mapstructure/v2 from 2.2.1 to 2.3.0 [#18404](https://github.com/vitessio/vitess/pull/18404)`
- `* Main: Bump Go Dependencies [#18686](https://github.com/vitessio/vitess/pull/18686)`
- `* Upgrade the Golang Dependencies [#18711](https://github.com/vitessio/vitess/pull/18711)`
- `* Bump @babel/runtime from 7.26.0 to 7.27.6 in /web/vtadmin [#18467](https://github.com/vitessio/vitess/pull/18467)`
- `* Bump form-data from 4.0.1 to 4.0.4 in /web/vtadmin [#18473](https://github.com/vitessio/vitess/pull/18473)`
- `* Bump vite from 4.5.9 to 4.5.14 in /web/vtadmin [#18485](https://github.com/vitessio/vitess/pull/18485)`
- `` * Deprecate `github.com/mitchellh/mapstructure` library [#18687](https://github.com/vitessio/vitess/pull/18687) ``

## Other changes (110)

- VTTablet state metric gains `type`, `keyspace`, and `shard` labels
- SQL statement splitter handles `DEFINER` clauses in `CREATE PROCEDURE`
- Improved splitting of `CREATE PROCEDURE` statements with `DEFINER` clauses
- Stored procedure parsing no longer converts `START TRANSACTION` to `BEGIN`
- Window functions pushed down to the target shard
- More aggressive merging of UNION queries
- Unnecessary derived-table wrapping avoided for UNION queries
- Read-only multi-shard transactions in `SINGLE` mode
- Prepared statements now support runtime-dependent query plans
- Recovery stats now include keyspace and shard labels
- AI-use disclosure added to the pull request template
- MySQL Shell restore capability check corrected
- CI builds use event payload for draft state
- Local-example tests pass `TOPO` to Docker and omit Consul
- golang-ci lint script accepts versions with or without a `v` prefix
- Canceled vreplication migration entries retained
- Subquery merging restored in the planner
- `START TRANSACTION` preserved in `CREATE PROCEDURE` bodies
- Database qualifiers removed from tablet queries
- NamedLock tests added and passing for zk2 and Consul
- Expanded SwitchWrites progress and error logging
- VTCombo test helper selects the version-appropriate `--vschema-ddl-authorized-users` variant
- No post-return callbacks from `*vstreamManager.VStream`
- Immediate return from `*vstreamManager.VStream` on callback errors
- Sequence counters reset when traffic switches
- Go-version upgrade workflow includes the `release-22.0` branch
- CI workflows switch to CNCF Oracle VM runners
- `depguard` restrictions for library packages
- Go dependency and Go version workflows update `peter-evans/create-pull-request` to v7
- CI workflows now use MySQL 8.4 by default
- Protobuf randomness hack removed; golden tests now compare structure
- CI workflows updated to `mysql-apt-config_0.8.35-1_all.deb`
- Vitess Java package publishing moved to Central
- Flaky vdiff2 end-to-end test moved to its own workflow
- `TestMigrateSharded` skipped in CI
- v22.0.0 release notes copy edited
- `MAINTAINERS.md` contact email updated
- Release-cycle pull-request analysis tooling added
- vmg moved to the past-maintainers list
- `@deepthi` moved to emeritus
- Manan Gupta’s maintainer email updated
- Manan Gupta moved to past maintainers
- Vitess maintainer list updated
- `CODEOWNERS` ownership entries updated
- Summary note on `CREATE PROCEDURE` support through vtgate for unsharded keyspaces
- `go-upgrade` detects Go patch-version updates
- CI workflow filters moved into event configuration
- Querylog output gains an `EmitReason` column
- Command-line flags migrated to dash-separated names
- VTTablet ACL reload success log formatting fixed
- Migration UUID added to Online DDL cut-over logs
- More `UNION` queries merged and tuple bind variables de-duplicated
- Parsed comments retained in merged subqueries
- Vitess maps MySQL 9.x to the MySQL 8.2 flavor
- Wildcard database-name checks added to external-tablet `SwitchTraffic` permissions
- External tablet permission-check logging corrected
- VReplication copy-phase history-list limit raised to `10MB` by default
- 51 `vtcombo` flags migrated to dashed names
- Context added to `heartbeatWriter` errors
- Smart connection pool semaphore switched to channels
- `EmergencyReparentShard` candidate selection uses SQL-thread progress
- MySQL Shell backup restore log corrected for REDO_LOG re-enablement
- CI workflows use `!cancelled()` instead of `always()`
- CI workflows upgrade `actions/setup-go` to v5.5.0
- GitHub Actions dependency installation skips `man-db` updates
- Invalid comment lines removed from `vttablet-up.sh`
- Code Coverage workflow filters push-event runs
- Unit-test runner runs packages in parallel
- Clarified `NewHealthCheck` filter documentation
- Local examples use dashed flags
- asthelpergen design documentation
- VTOrc renames `isClusterWideRecovery` to `isShardWideRecovery`
- VTOrc replication analysis drops `ClusterDetails`
- VTOrc shard primary timestamp handling updated
- VTOrc keyspace and shard refreshes use the errgroup-derived context
- VTOrc reads tablet type fields with `GetInt32`
- GTID set updates and `vstreamer` formatting optimized
- `vstreamer` `CurrentTime` generation per GTID group
- VTorc GTID-set handling switches to `replication.Mysql56GTIDSet`
- Multi-query protocol flag handling fixed
- Error-free handling of `None` opcode DML queries
- v23 upgrading notes reference `vitess/lite:v23.0.0`
- v22.0.0-RC1 release notes added to the main changelog
- v21.0.4 release notes added to the main changelog
- Added `v20.0.7` release notes to the main changelog
- v22.0.0-RC2 release notes added to the main changelog
- `v22.0.0-RC3` release notes added to the main changelog
- v22.0.0 release notes added to the main changelog
- `v22.0.1` release notes added to the main changelog
- v21.0.5 release notes added to the main changelog
- v20.0.8 release notes added to the main changelog
- Code freeze recorded for `v23.0.0-RC1`
- Vitess `v23.0.0-RC1` release
- Vitess v23.0.0-RC2 release recorded
- Version bumped to v23.0.0-SNAPSHOT after v23.0.0-RC2
- Release notes updated
- Longer timeout for `TestTrackerNoLock` and basic check for `TestCreateLookupVindexMultipleCreate`
- `TestVStreamsMetricsErrors` handles race-dependent outcomes
- MySQL directory collisions fixed in end-to-end tests
- VExplain test sets the `_binary` prefix only with vtgate v23 or later
- `TestQueryTimeoutWithShardTargeting` accepts varied error messages
- Tests for empty JSON array inserts
- More flexible `TestOverallQueryTimeout` error matching
- MySQL 8.0 support for foreign-key end-to-end tests
- `FileBasedConfigLoader` tests now match expected error messages exactly
- VStream metrics test flakiness fixed
- Atomic heartbeat counter in `TestStreamRowsHeartbeat`
- VDiff2 end-to-end test waits for VDiff to start or complete before stopping
- Semi-sync monitor test races fixed
- Smart connection pool test timeout increased

---

Summaries are produced by AI, so they can be wrong, and they are provided without warranty.
