Repository navigation
fix: combine durable activity, stable memory and native content - #236
Merged
Merged
Conversation
…sue215-durable-public-activity
This was referenced Oct 20, 2026
lyyzka
marked this pull request as ready for review
October 20, 2026 14:55
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Memory lookup no longer registers evolution benchmarks on ordinary reads, so an unrelated evolution-table lock no longer stalls the lookup path. The combined release also preserves durable authorized memory, reconnect/history activity and native rich-content boundaries. Contributor commits from PR225, PR224→232 and PR223→231→233 are retained.
Deployed and accepted on 2026-10-20 at14:36:35 UTC: exact application
d1075e35e16721979b484c2d9e33eeb56468ce9d, formally published LingxiOS3.5.8, native schema13/protocol11 and product migration48. Both Webs and Workers are healthy, zero restarts/OOM, inspected policies unchanged;18 final checks and independent reconciliation pass. Four HTTPS routes and all three entry assets match the qualified image. OCI indexsha256:697d5a3e3d34629e603c120cd068abde3a2add611f583d5444d77a947887bf69; linux/amd64 manifestsha256:1c437ff4f2b2f36055fddcf179460963d9f95147406989b48355a0020e4bf862.Qualification includes70/70 PostgreSQL/native Worker/HTTP/SSE/business checks,34/34 migrations,69/69 web and118/118 server checks,38/38 product and7/7 native-editor official-engine browser checks,21/21 cancellation teardown, and an independent formal356 backup/restore→358 rehearsal. Exact d107 CI39412540194 and publication39413726289 passed; each full integration run had340 passes,0 failures,13 explicit skips. Real preproduction GLM5.2+bge-m3 qualification covered9 behaviors/39 measured-usage HTTP200 calls.
The controlled1200ms unrelated evolution-lock comparison was1204.32ms on historical3.5.3 versus28.69ms on formal3.5.8, under an unchanged500ms budget. Ordinary reads perform zero evolution registrations. The0/8/1024-document first/warm/concurrent samples are not general production percentiles; the earlier207-second Abort remains unattributed. Models and databases are not asserted as unproven root causes.
The final production attempt used a new173-table backup/isolated restore with equal counts and provenance before migrations47/48. The earlier failed25-minute checkout incident and compatible recovery remain retained. Online native host-file preparation removed repeated Git from the maintenance path. A later WebB mirror522 stopped recovery safely; a separately reviewed continuation restored only remaining services. All27 final terminal updates are accounted for:26 successes and that explicit failure. Two verifier assumptions (Docker ImageId vs OCI config digest, and the frozen Tini entrypoint) were corrected with failed artifacts retained, without restarting services to satisfy checks.
This successful attempt's public readiness samples span199.732seconds from first failure to recovery; WorkerA/B became healthy later at14:32:03/14:33:07 UTC. HTTP readiness is not authenticated conversation availability. Authenticated production business/provider smoke was not run; no credential was created/copied. Startup isolation checks are inferred from exact shipped startup/readiness, not a complete production fault-injection matrix. The paused fixed3.5.3 soak remains untouched.
Accepted production evidence, exact receipts, backup, availability and independent review · Formal qualification.
Related #196, #213–#219. PR237 and #229 remain separate: actual final-image savings, controlled cold/warm builds and peak-disk measurements are not claimed complete here. Host-file mode remains active; the legacy broad rollout stays blocked pending its separately qualified same-schema path. Do not start old native12 Workers against schema13.