[CASSANDRA-17235] Add diagnostic events for unavailable errors

• Type: New Feature
• Status: Patch Available
• Priority: Normal
• Resolution: Unresolved
• Component/s: Observability/JMX
• Change Category: Operability
• Complexity: Normal
• Platform: All
• Assignee: Stefan Podkowinski
• Reporter: Stefan Podkowinski

Authors

Stefan Podkowinski

Reviewers

Marcus Eriksson

• Created: 04/Jan/22 13:57
• Updated: 05/Jan/22 10:13

Description

Errors from unavailables during query execution lack important details on the context why the error occurred. Emitting diagnostic events on unavailables will allow you to attach to the instance during troubleshooting and observe the type of operation (what kind of reads/writes), keyspace replication settings and down nodes as seen by the instance throwing the error.