chore(client): remove auto metadata-backfill hook; run backfill manually

[divyasinghds]

What

Removes the automatic metadata-backfill runner (the post-upgrade Helm hook added in #380). The backfill will be run manually as a one-off Job instead of firing on every chart upgrade.

Deleted:

• client/templates/metadata-backfill-hook.yaml — the post-upgrade hook Job
• client/tests/metadata_backfill_hook_test.yaml — its helm-unittest
• metadataBackfill block in client/values.yaml (+ comment header)
• metadataBackfill property in client/values.schema.json

Why

The auto-hook coupled a one-time, best-effort data migration to the chart's upgrade lifecycle, which repeatedly surfaced timing/--timeout edge cases in review. Since the backfill is genuinely a one-time operation per cluster, running it manually is simpler and removes that coupling entirely. The runner itself (tracebloc-backfill entrypoint, data-ingestors #393/#395) is unchanged and stays the mechanism — we just invoke it on demand.

Manual run (replaces the hook)

Run the ingestor image's entrypoint as a standalone Job in the client namespace:

apiVersion: batch/v1
kind: Job
metadata:
  name: metadata-backfill-manual
spec:
  backoffLimit: 0
  activeDeadlineSeconds: 1800
  template:
    spec:
      restartPolicy: Never
      containers:
        - name: backfill
          image: ghcr.io/tracebloc/ingestor:0.7.7
          command: ["tracebloc-backfill"]
          env:
            - {name: CLIENT_ID,       valueFrom: {secretKeyRef: {name: tracebloc-secrets, key: CLIENT_ID}}}
            - {name: CLIENT_PASSWORD, valueFrom: {secretKeyRef: {name: tracebloc-secrets, key: CLIENT_PASSWORD}}}
            - {name: MYSQL_HOST,  value: "mysql-client"}
            - {name: CLIENT_ENV,  value: "dev"}
            - {name: LOG_LEVEL,   value: "INFO"}
            - {name: HOME,        value: "/tmp"}

Verified on dev (tb-client-dev-templates / tracebloc-templates): ok=98, skipped_current=26, not_found=26, errors=0. The sweep is idempotent, so re-running is safe.

Testing

• helm lint client passes.
• helm template renders 0 backfill objects.
• values.yaml / values.schema.json validate.

🤖 Generated with Claude Code

---

Note

Low Risk
Chart-only removal of an optional hook; no changes to ingestor runtime, jobs-manager, or auth paths. Upgrades no longer run backfill automatically—clusters that still need migration must run it manually once.

Overview
Removes the automatic federated-alignment metadata backfill from the client Helm chart. Chart upgrades no longer spawn a post-upgrade hook Job that ran tracebloc-backfill on the ingestor image.

Deleted: metadata-backfill-hook.yaml, its helm-unittest suite, and the metadataBackfill values block (including enabled, backfillTimeoutSeconds, and resources) from values.yaml and values.schema.json.

The tracebloc.ingestorDigest helper comment is updated so ingestor image resolution is documented for jobs-manager ingestion Jobs and manual backfill Jobs only—not the removed hook.

Operators run the same tracebloc-backfill entrypoint as a one-off Job when needed; the sweep stays idempotent.

^{Reviewed by Cursor Bugbot for commit 853ef89. Bugbot is set up for automated code reviews on this repo. Configure here.}

[LukasWodka]

👋 Heads-up — Code review queue is at 44 / 30

Above the WIP limit. The team convention is to review existing PRs before opening new work.

Open PRs currently in Code review (oldest first):

• averaging-service#207 — Promote develop → staging · author: @shujaatTracebloc · no reviewer assigned
• backend#1113 — perf(serving): gthread workers + fix CONN_MAX_AGE/CONN_HEALTH_CHECKS env parsing (#976) · author: @aptracebloc · reviewer: @saadqbal, @saqlainsyed007
• backend#1201 — docs(rfc): Data spaces & federated dataset composition (draft) · author: @LukasWodka · reviewer: @saadqbal
• backend#1217 — docs(rfc): Dataset integrity & score reproducibility (draft) · author: @LukasWodka · reviewer: @saadqbal
• backend#1228 — Promote develop → staging · author: @shujaatTracebloc · reviewer: @saadqbal
• backend#1244 — fix(experiments): surface save-time failures as 5xx + guard new NOT NULL columns · author: @LukasWodka · reviewer: @saqlainsyed007, @aptracebloc
• backend#1251 — docs(bugbot): add .cursor/BUGBOT.md project context (#930) · author: @LukasWodka · reviewer: @waqaskhanroghani
• backend#1254 — docs(pr-template): owner-qualify cross-repo closing keywords · author: @LukasWodka · reviewer: @waqaskhanroghani
• backend#1259 — docs(rfc): qualified RFC identifiers + org-wide index · author: @LukasWodka · reviewer: @saadqbal
• backend#1261 — RFC-BACKEND-0008: board & process reset (decision RFC — for markup) · author: @LukasWodka · reviewer: @saadqbal

Pull from review before opening new work. (This is a nudge from the kanban WIP check, not a block.)

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 05627ea. Configure here.}