Details
-
Feature Request
-
Resolution: Done
-
Major
-
2.4 GA
-
Not Started
-
Not Started
-
Not Started
-
Not Started
-
Not Started
-
Not Started
-
0
-
0%
Description
Sometimes in on-premise deployments there can be intermittent network drops between pods. This can cause issues with the pods that communicate with redis (backend-listener, system-reque, etc).
The issue presents itself as either timeouts or connection reset errors.
Connection timed out (Redis::TimeoutError)
Redis::ConnectionError - Connection lost (ECONNRESET)
It would be useful if the clients that communicate with Redis retried the connection.
When present, this issue will also cause a significant amount Admin API calls to fail completely due to the redis connection error. If the connection could be retried before failure it may prevent this.
Attachments
Issue Links
- is related to
-
THREESCALE-608 backend-listener fails to connect backend-redis
- Closed
- relates to
-
THREESCALE-4180 Restart sidekiq automatically when jobs are failing
- Closed