# Internal Only Resolved Cluster Identity

<!-- auto-oncall-optimization-record
{
  "schemaVersion": "auto-oncall-optimization-record-v1",
  "optimizationId": "internal-only-resolved-cluster-2026-08-14",
  "problems": [
    {
      "problemId": "internal-only-omits-resolved-wetech-cluster",
      "summary": "APID-29343 successfully queried the resolved Wetech Dolphinscheduler target, but Internal Only omitted the region and cluster identity, leaving Support unable to tell which cluster the evidence came from.",
      "status": "optimized_verified",
      "applicability": "Wetech diagnostic response pairs whose Scenario Router resolved an exact region and cluster and reached Provider acquisition.",
      "issueKeys": ["APID-29343"],
      "sourceRefs": [
        "https://tidb.atlassian.net/browse/APID-29343#problem:internal-only-resolved-cluster-omitted"
      ],
      "signalTextIncludesAny": [
        "support cannot find the cluster",
        "Internal Only omitted cluster identity",
        "resolved Wetech target"
      ],
      "observedThrough": "2026-08-14",
      "effectiveAt": "2026-08-14T03:59:54Z",
      "verifiedAt": "2026-08-14T04:18:55Z",
      "attemptedAction": "Require Scenario Router to place the exact resolved Wetech region and cluster name in Internal Only and require Final Response to preserve it without exposing machine target references or moving it into the customer message.",
      "evidenceUrl": "https://tidb.atlassian.net/browse/APID-29343?focusedCommentId=23563769"
    }
  ]
}
-->

- Optimization ID: `internal-only-resolved-cluster-2026-08-14`
- Owner: Scenario Router and Final Response audience projection
- Source reference: `APID-29343`, internal comment `23563769`
- Implementation status: merged and deployed to staging
- Effective status: `optimized_verified`

## Bounded problem

APID-29343 resolved the Wetech route to `asia-prod / Dolphinscheduler` and
successfully acquired TiKV metrics from that target. The generated `Internal
Only` section described the evidence gap and next Support checks but omitted
the resolved region and cluster name. Support therefore could not tell which
actual cluster the evidence represented.

## Implemented change

- Require Scenario Router to add one compact Support-readable queried-cluster
  line when the Wetech route has an exact region and entity.
- Require Final Response to preserve that line only in
  `internalHandoffMarkdown`.
- Keep machine-form `wetech:` references, datasource UIDs, sessions,
  credentials, and routing state out of both response messages.

## Verification evidence

The focused Scenario Router, Final Response, and support-reply contract tests
passed 65/65. The complete local TypeScript build and test suite passed
819/819. Source PR
[`tidbcloud/aisupport#907`](https://github.com/tidbcloud/aisupport/pull/907)
merged as `baf9cfbd8e91b5eb8ec8125e634be59b5ac1f8bd`. The staging
deployment workflow and generated
[`tidbcloud/app-delivery#77282`](https://github.com/tidbcloud/app-delivery/pull/77282)
completed, and the live Ready Pod used that exact image with zero restarts.

Authorized write replay `0339087b-da95-47bf-9db7-d73b0fed2b88` completed with
`decision=internal_only`. Jira comment `23563769` was updated in place and read
back as `public=false`; its Internal Only section contains
`查询集群：asia-prod / Dolphinscheduler`. The customer reply contains none of
`asia-prod`, `Dolphinscheduler`, or a machine `wetech:` target reference.

## Unresolved gaps

None for this bounded cluster-identity problem. The replay's separate
diagnostic evidence result remained degraded because a Scenario Diagnostic
Agent exhausted its budget; that did not alter the resolved Wetech target or
the verified audience projection.

## Next validation action

Monitor later Wetech responses for the same cluster line and treat any omission
from Internal Only, or appearance in the customer reply, as a regression.