Details
-
Bug
-
Resolution: Done
-
Major
-
None
-
False
-
None
-
False
-
NEW
-
NEW
-
Before this update, the LokiStack route configuration caused queries running longer than 30 seconds to timeout. With this update, the LokiStack global/per-tenant queryTimeout setting affects the Route timeout setting resolves the issue.
-
Bug Fix
-
Log Storage - Sprint 236, Log Storage - Sprint 237
-
Important
Description
Description of problem:
Needed to review and fix all the points for being able to see the logs from the OCP Console from Observe > Logs when the backend storage is Loki.
Loki is failing from the Console to return results when the query is exceeding 30 seconds. In the past, some bugs were opened and also was opened an RFE to OCP Console:
- bug LOG-3498 where a hardcoded timeout was in the logging-view-plugin.
- timeout to 30 seconds coming from the Console: https://issues.redhat.com/browse/RFE-3775
Also, after commenting with engineering, the timeouts in Loki, have different/inconsistent values:
- logging-view-plugin: 30s
- lokistack-gateway: 10m
- loki: 1m
Needed to make the product consistent in the timeout values and even if possible, configurable directly from the end-user/admin of the OCP Cluster being able to query logs without any problem.
Version-Release number of selected component (if applicable):
RHOL 5.6 using Loki as backend storage
How reproducible:
As reference LOG-3498 where QE was failing
Steps to Reproduce:
Check LOG-3498 for reference and QE tests.
Actual results:
Failing when the query is longer than 30s. Not difficult to hit it when asking 1 day or 2 days of logs
Expected results:
Able to see logs in the OCP Console > Observe > Logs when the query is taking longer than 30 seconds.
