We are continuing to monitor for any further issues.
Posted Aug 27, 2025 - 17:53 PDT
Monitoring
During a graph database migration to improve search latency, nodes in cluster ran out of file descriptors. We quickly identified the issue and implemented a fix. Nodes took some time to recover and rejoin the cluster. We will be running a post-mortem to understand and address shortcomings in our systems and processes around scaling our graph database cluster, including maintenance and development processes.