fix(selfhost): add the ams_signals conflict key so Postgres AMS ingest stops throwing

[kai392]

Summary

• Add ams_signals to REPLACE_CONFLICT_KEYS in src/selfhost/pg-dialect.ts. Without it, translateInsertOr threw pg_dialect: INSERT OR REPLACE into 'ams_signals' has no known conflict key on every self-host Postgres deployment, so the first AMS telemetry-ingest write (src/ams/ingest.ts, reached via POST /v1/ams/ingest) failed the request outright.

The key-shape evaluation the issue asked for — the 2-column key is correct here

Per the issue's second branch: the orb_signals collision reasoning does not apply to ams_signals, so no migration is added and the entry matches the table's existing constraint.

• Both tables are fed by the same exporter, packages/loopover-miner/lib/orb-export.ts (DEFAULT_AMS_COLLECTOR_URL = ".../v1/ams/ingest"), which derives:
  • repoHash = hmacAnonymize(repoFullName, secret)
  • prHash = hmacAnonymize(${repoFullName}:${prNumber}, secret)
• The repo is therefore inside the pr_hash input: within one instance_id (one secret), a pr_hash already identifies (repo, PR), and repo_hash is functionally determined by it. Two different repos can never collide on (instance_id, pr_hash) — exactly the "pr_hash's hash input already disambiguates by repo" case the issue names.
• Widening would also be actively wrong to ship alone: the map must name the table's real constraint (UNIQUE (instance_id, pr_hash), migrations/0148_ams_signals.sql), or Postgres rejects the generated ON CONFLICT with "no unique or exclusion constraint matching the ON CONFLICT specification". A 3-column target would need a table recreate for zero uniqueness gain.
• A comment at the entry records this reasoning (and the one condition that would reopen it: pr_hash ceasing to be repo-scoped at the producer), so it isn't silently re-litigated later.

Test plan

• test/unit/selfhost-pg-dialect.test.ts — new regression test translating the verbatim statement src/ams/ingest.ts issues: no throw, ON CONFLICT (instance_id, pr_hash) DO UPDATE SET, every non-key column upserted (repo_hash, decision, reason_bucket, closed_at, received_at), and both key columns excluded from the SET list
• Suite green (23/23); local coverage on src/selfhost/pg-dialect.ts: 41/41 lines, 14/14 branches
• npm run typecheck clean
• No migration and no src/ams/ingest.ts change needed (see the evaluation above), so db:migrations:check/db:schema-drift:check are unaffected
• CI validate

Closes #8382

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.56%. Comparing base (d8f706c) to head (94a7e63).
⚠️ Report is 38 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8398      +/-   ##
==========================================
- Coverage   92.12%   89.56%   -2.57%     
==========================================
  Files         782       98     -684     
  Lines       78531    22663   -55868     
  Branches    23720     3872   -19848     
==========================================
- Hits        72347    20298   -52049     
+ Misses       5062     2187    -2875     
+ Partials     1122      178     -944     

Flag	Coverage Δ
shard-1	60.97% <ø> (+1.02%)	⬆️
shard-2	51.21% <ø> (+4.00%)	⬆️
shard-3	100.00% <ø> (+46.47%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/selfhost/pg-dialect.ts	100.00% <ø> (ø)

... and 684 files with indirect coverage changes

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-24 11:31:02 UTC}

2 files · 1 AI reviewer · no blockers · readiness 95/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This PR adds a single map entry `ams_signals: ["instance_id", "pr_hash"]` to `REPLACE_CONFLICT_KEYS` in pg-dialect.ts, fixing a real bug where every self-host Postgres AMS ingest write threw `no known conflict key`. The chosen 2-column key correctly matches the table's actual `UNIQUE (instance_id, pr_hash)` constraint from migration 0148 (verified in the provided migration file), unlike orb_signals' 3-column key which matches a different constraint shape. The regression test translates the verbatim INSERT statement from ams/ingest.ts and asserts the exact ON CONFLICT target and SET list, which is a real, non-fabricated test of the fixed path.

Nits — 4 non-blocking

• The inline comment in pg-dialect.ts is quite long (10 lines) relative to the one-line map entry it documents; could be trimmed or moved to a doc reference, though the reasoning is genuinely useful for future maintainers.
• The regression test hard-codes column names/order matching ams/ingest.ts's INSERT statement; if that statement's column list changes, the test will need manual updating (no shared source of truth), same pattern as other tests in this file so consistent with convention.
• Consider a short comment-length convention check across REPLACE_CONFLICT_KEYS entries so future additions don't grow entries this large; not necessary here since the reasoning materially prevents a re-litigation risk called out in the PR description.
• The integration test file (selfhost-pg.test.ts) doesn't appear to test ams_signals against a real Postgres instance the way it does for system_flags — consider adding a real-PG integration case for parity, though the unit test already exercises the translation logic directly.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#8382
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (1 linked issue).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 107 registered-repo PR(s), 58 merged, 4 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	kai392; Gittensor profile; 107 PR(s), 4 issue(s).
Improvement	✅ Minor	risk: clean · value: minor · LLM: moderate

Linked issue satisfaction

Addressed
The PR adds ams_signals to REPLACE_CONFLICT_KEYS with the correct (instance_id, pr_hash) key matching the existing migration constraint, and explicitly performs and documents the collision-risk evaluation the issue requested (concluding widening is unnecessary since repo is embedded in pr_hash), plus adds a regression test on the verbatim ingest statement.

Review context

• Author: kai392
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: JavaScript, Kotlin, Perl, Python, TypeScript, Vue
• Official Gittensor activity: 107 PR(s), 4 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Start here: Triage stale or unlinked PRs.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by LoopOver, a quiet PR intelligence layer for OSS maintainers.

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.