Exact replacement service/query results

Product e078ba23b543126011a2136dd1f7e98ca65370eb, package source 28bcdf7ac833ab634d3b130b85cb2a0aa47eb7a8, runtime c14092e89f86cb8e35d57b6aa3fafd646bfa0863.

All six full main and four full query cells completed with actual exit 0, matching immutable package/source identities, confirmed cleanup and stopped outer resource groups. Independent canonical closeout exited 0 at 2026-10-18T05:04:24.453109+00:00. Private terminal index SHA-256 c19ec8f31b97d3405d79d709ad95eacf4e9d205c2fee998f55d427b36aa28cc1; canonical completion SHA-256 595a62685f68c2de58b096c9d82ae28f0fdf7fc30fda4bb8de5a733a6b91d16e.

DEB SHA-256 2a555bec6334fde6774d21c414d7d42e041c39ca170e3a30fd3676e134b9f593; RPM SHA-256 12c03f92e96c7d1a848a5198f017c21a1e17d485cdfca7feb8464c3a425971c1.

Cell	Gates	Maximum registered p99, seconds
replacement-tls-01-service-main-t10-s1-01	13	0.022546551
replacement-tls-01-service-main-t10-s2-01	13	0.022718024
replacement-tls-01-service-main-t10-s3-01	13	0.022693815
replacement-tls-01-service-main-t100-s1-01	13	0.112976906
replacement-tls-01-service-main-t100-s2-01	13	0.109098274
replacement-tls-01-service-main-t100-s3-01	13	0.112012384
replacement-tls-01-service-query-segment-s1-01	8	1.220481050
replacement-tls-01-service-query-segment-s2-01	8	1.221866073
replacement-tls-01-service-query-segment-s3-02	8	1.209791000
replacement-tls-01-service-query-journal-s1-01	8	1.180263220

Main latency is combined measured creation/acceptance-to-ACK p99 for that cell; query latency is the maximum of its ten kind/phase p99 values. All six main cells kept their original 135-second exposure, tier/seed combinations and clearing-v2 rule. Each full query retained 200 independent answer verdicts, 100 sources, 10,000 batches, 500,000 logs and 500,000 metrics; canonical closeout separately decoded each source archive’s 1,000 traces/3,000 spans and exact root/two-child parentage.

Positive closeout exited 0. A copied queue changed only cells[0].summary_sha256; the negative exited 1 with exactly closeout cell receipt binding mismatch. Canonical closeout then exited 0. All three resource groups stopped. Maximum observed DATA usage was 80,572,618,440 bytes, below the 95,000,000,000-byte stop line.

The owner pause interrupted Segment seed 3 trial -01; its exit-2 wrapper, raw files, scratch and failure evidence remain preserved and excluded. The append-only continuation reused the original eight full results and three bootstrap records after validation, then ran fresh Segment seed 3 -02 and the original journal trial. Controller/closeout variants changed only administrative paths; the original semantic graders and 13-case population stayed unchanged.

These are bounded synthetic service/query results. They do not establish full soak endurance, packaged browser/PWA, cross-family/outage, owner walkthrough or release acceptance. Sampled-v1 and integrated backlog diagnostics remain visible; no claim of stationary or universally nonincreasing mean backlog is made.

The coordinator independently checked the actual completion identity, ten-cell population, table metrics, primary wrapper hashes and outer stops before publication. No release or tag is claimed.

Registered service/query protocol. Private primary records are results/service-query-replacement-tls-01/service-query-actual-terminal-evidence-02.json and completion-02.json; their hashes above bind the original raw inventories and per-case commands.