Timeouts reported on our East Coast Cluster

22:27 UTC

We are investigating reports of timeouts on our AWS East Coast cluster (“Julep”) that started after 21:00 UTC and are currently affecting some Heroku customers. We’ll report more details here as our investigation unearths them.

Update 00:56 We’ve identified a failed disk on one of the servers powering this cluster. Users with existing databases who submit requests to create the same database again will receive timeouts while this server is down.

Update 02:47 A replacement server is online and databases are now being restored to nominal replication levels.