Apr 18, 14:42 UTCResolved. Batch queue drained. All endpoints returning normal latency. Root cause: memory pressure on worker pool eu-west-1b.
Apr 18, 13:11 UTCMonitoring. Deployed hotfix to worker pool. Processing times improving, queue draining at ~400 jobs/min.
Apr 18, 12:58 UTCInvestigating. Identified memory leak in batch worker v2.3.1. Rollback to v2.3.0 initiated across eu-west-1b nodes.
Apr 18, 12:31 UTCIdentified. Batch endpoints showing P99 > 4000ms. Data API unaffected. Team investigating.