Degraded Performance on Database

Incident Report for PostHog

Resolved

Part counts are back to normal and replicas are up to date! Happy Hoggin!

Posted Jan 29, 2021 - 19:18 UTC

Monitoring

The issue with performance was related to a spike in parts for one of our reporting tables dramatically slowing down query times. This table has been replaced with a table with much fewer parts. It will take some time for replication to occur across the cluster.

Posted Jan 29, 2021 - 18:36 UTC

Investigating

We are currently investigating why a table on our OLAP database has significantly increased number of parts causing backups of queries. Data ingestion is not effected but the app and dashboards may be unstable.

Posted Jan 29, 2021 - 17:33 UTC

This incident affected: US Cloud 🇺🇸 (App).