perf: fix slow /packets?groupByHash=true on large stores#2
Open
perf: fix slow /packets?groupByHash=true on large stores#2
Conversation
efiten
force-pushed
the
fix/grouped-packets-perf
branch
from
38e1e38 to
f80974b
Compare
- Add LatestSeen field to StoreTx, maintained in all three observation write paths (load, real-time ingest, poll). Eliminates the per-packet observation scan that was O(total_packets * avg_observations). - Build grouped packet maps under read lock (correct), sort the local copy outside the lock (avoids holding lock during O(n log n) sort). - Cache the full sorted result for 3 seconds keyed by filter params. Repeated requests within the TTL return instantly without re-sorting. Fixes /packets?limit=50000&groupByHash=true taking 16s on large stores. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
GetChannels was iterating all payload-type-5 packets and JSON-unmarshaling each one while holding s.mu.RLock(), blocking all concurrent reads. On stores with many channel messages this caused /api/channels to take 13s+. - Copy only the needed fields under the read lock, release before unmarshal - Cache the result for 15 seconds keyed by region param - Invalidate cache on new packet ingestion Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
efiten
force-pushed
the
fix/grouped-packets-perf
branch
from
f80974b to
3bf1df5
Compare
…channel cache - TestLatestSeenMaintained: verifies StoreTx.LatestSeen is set >= FirstSeen and >= all observation timestamps after store load - TestQueryGroupedPacketsSortedByLatest: verifies packets with more-recent observations sort before packets with newer first_seen but older observations - TestQueryGroupedPacketsCacheReturnsConsistentResult: verifies cache returns consistent total and ordering on back-to-back calls - TestGetChannelsCacheReturnsConsistentResult: verifies GetChannels cache returns same channel names on repeated calls - TestGetChannelsNotBlockedByLargeLock: verifies GetChannels returns correct data (channel name, messageCount) after lock-copy-unmarshal refactor Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
efiten
force-pushed
the
master
branch
2 times, most recently
from
41bf3b9 to
f0fc940
Compare
efiten
force-pushed
the
master
branch
2 times, most recently
from
493d9e8 to
568de4b
Compare
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
1 participant
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.
Summary
LatestSeenfield toStoreTx, maintained in all three observation write paths (load, real-time ingest, poll) — eliminates the O(n×observations) scan that was happening per packet at query timeFixes
/packets?limit=50000&groupByHash=truetaking 16s on large stores.Test plan
./deploy-staging.sh fix/grouped-packets-perf[SLOW API]log in browser console is gone🤖 Generated with Claude Code