jordan fe8d0c87e7 harden: restore CI verification, remove four wire-level fabrications, instrument the 401 path

Implements tmp/tidaldb-fleet-hardening (20 planned tasks + 2 found by measurement).

Ring 0 — restore verification. .woodpecker.yaml step pods ran at the namespace
default of 1500m/2Gi, which OOMKilled a prior pipeline and starved the release
gate past its budget. Both push-path steps now declare
backend_options.kubernetes.resources as two YAML anchors declared once on their
first consuming step. The values are CALIBRATED against measured free node
capacity, not against the LimitRange max: `requests: cpu 2` (this roadmap's
original figure) fits on NO node and would sit Pending forever, because
`ci-build-bounds` grants permission and the nodes supply capacity, and those are
not the same thing.

The `nightly` cron described in this file for 216 days was never created, so
tier-3 chaos, the fault classes, mTLS and the PITR test produced exactly zero
signal while reading like standing coverage. nightly-chaos and
nightly-security-ops now alias the anchors and have budgets matching the gate
(their 120/90 were TIGHTER on the same runner, so they would have failed
nightly for a budget reason, not a correctness one). nightly-soak is REMOVED,
not scheduled: it drives 1000 rps for 600s gating on p99 <= 250ms, and the best
node has 1700m free CPU, so it would fail on starvation rather than regression —
manufacturing a nightly false alarm. Its commands move verbatim to
docs/runbooks/nightly-soak.md.

Ring 1 — four fabrications removed from the wire.
- scatter_merge sorted and truncated without re-stamping rank, so /feed and
  /search returned 1,1,2 under full placement. Reuses merge_cross_shard's
  existing stamp; asserted on BOTH the multi-group merge path and the
  single-group [only] fast path that bypasses it.
- aggregate_region_row's None arm invented `applied_events: 0` plus a deficit
  derived from it. applied_events/lag_events are now Option<u64>, null on the
  wire. leader_last_seq was also unwrap_or(0), so a node that could not reach
  the LEADER computed 0 - applied = 0 for every region and reported a converged
  cluster it had never measured — a fabrication pointing the dangerous way.
- tidalctl inferred NO REPORT from `applied == 0 && lag > 0`. That heuristic was
  actively hiding the PVC-wipe shape: a measured zero with a real deficit
  rendered as "no report" instead of BEHIND. Now read off the wire; converged
  exits 0, partitioned still exits nonzero.
- /sharded/* answered 201/204 for single-copy writes with nothing anywhere
  saying so. Now requires `x-tidal-ack: local`, rejecting with 400 via the
  existing invalid_input path. Six call sites migrated, not the two this
  roadmap predicted — including docs/runbooks/cluster.md §16.3, which told
  operators to run a quorum-write probe via POST /sharded/items. That probe
  cannot verify quorum: the surface applies locally with no WAL append. It was
  used as the safety check between every step of a staged deploy earlier today.

Ring 2 — observability. JSON_LOGS was already implemented and the deployment
simply never asked for it; the StatefulSet now sets it, plus
TIDAL_SERVICE_NAME=tidaldb because enabling it silently renames the
VictoriaLogs `service` stream field and would have blinded every query keyed on
it. Adds tidaldb_usearch_replicated_vectors_total, incremented on BOTH the
origin (wal_blob_first -> Ok(Some)) and the follower apply path — counting only
the origin would mean each vector lands on exactly one node, replicas never
agree, and the alert built on it pages forever.

Found by measurement, not planned: the 401 path discarded every fact about
every rejection. Traefik has served 101,858 rejected requests to the public
ingress — 87.6% of all its traffic — with no record of who or why anywhere.
unauthorized_response now emits reason (missing_token vs invalid_token, the
distinction that separates a scanner from a rotation that missed a consumer)
and the forwarded client. The token is never logged.

Also: scripts/restore-fleet.sh --cluster started the soak monitor while
deliberately leaving its gate suspended, orphaning a watcher that has reported
"0/30 green nights" for 13 days. The pair now moves together. Doc-guard's
three-warning backlog is cleared with real backfill for M4/M6/M12.

Verified: fmt clean; clippy 5 crates 0 new warnings (74 vs 74 baseline,
counted in a detached worktree at HEAD); lib 2110 passed; cluster_sharding 5;
cluster_runbook 10; tidalctl 38; doc-guard 0 warnings. Playwright 32/34 with
the two remaining failures asserting the rank fix against the not-yet-rolled
image — they are the post-deploy proof.

2026-08-30 20:55:58 -06:00

2.8 KiB

Raw Blame History

m12p1 — Read-recall harness: measurement truth (✅ COMPLETE 2026-06-14)

Landed in bb21e69. Changelog: CHANGELOG.md. Milestone index: README.md. Backfilled record.

Why this came first

G1 and G2 are numeric gates. Before m12p1 the project had no way to measure either: no pure ANN probe (every read path mixed profile scoring, fusion, and diversity into the result), and no ground truth to compare against. A gate you cannot measure is a gate you cannot pass — so the harness shipped before the tuning it exists to judge.

What shipped

A pure k-NN probe. TidalDb::vector_search_items(query, k, ef_search) returns raw HNSW nearest neighbours over the item content embedding slot with no profile scoring, fusion, or diversity, so the result measures index quality in isolation — the G2 metric. Exposed as POST /vector_search on the standalone router and on the multi-process region node, which merges by distance across hosted shard groups. A dimension-mismatched query is a 400, not a 500: caller error, not server fault.
The oracle. tidal-stress --verify-recall (tidal-stress/src/recall.rs) seeds the corpus with deterministic, id-keyed embeddings (reproducible with --skip-seed), holds a brute-force cosine ground truth in RAM, and ramps /vector_search probes open-loop with coordinated-omission correction.
A verdict, not a number dump. It reports per-stage true p99 and mean recall@k, plus the read-knee — the highest sustained QPS at which p99 ≤ target AND recall@k ≥ target both hold — with a machine-readable JSON summary (tidal-stress/src/summary.rs) and a --fail-on-knee PASS/FAIL exit code. Knobs: --recall-k, --recall-queries, --read-p99-target-ms, --recall-target, --recall-ef-search.
Repaired fabricated numbers. The p99 column in docs/profiling/social-scale.md was mean-as-p99 (the social bench is Criterion, which reports mean only). That column and the scale.rs / scale-baselines framing were corrected: every closed-loop number is now labelled an isolated per-op mean (regression tripwire), and only the open-loop harness signs off p99/recall SLOs.

Evidence

recall@10 = 0.9997 at 20k items / 1536-D against a real standalone server (HNSW M=16 / ef=400 / F16 vs brute-force cosine) — far above the 0.95 G2 target.
tidal/tests/m12p1_vector_recall.rs (3), tidal-server/tests/vector_search.rs (8).
Harness section in docs/profiling/scale-baselines.md.
A 1536-D HNSW-vs-brute recall@10 bench added to tidal/benches/vector.rs.

Carried constraint

The brute-force oracle needs roughly 6 GB RAM at 1M vectors, which is why the 100k/1M exit-gate runs execute the same harness on the k3s cluster rather than a laptop.

2.8 KiB Raw Blame History