-
Bug
-
Resolution: Done
-
Major
-
netobserv-1.6
-
None
-
False
-
None
-
False
-
When Loki is disabled, in the topology view of the Console plugin, when a group is set, and then the scope is changed so that the group becomes invalid, an error is displayed. Now the invalid group is removed, preventing the error to be displayed.
-
Bug Fix
-
-
-
NetObserv - Sprint 255
Description of problem:
In the topology view, when a group is set, and the scope is changed so that that group becomes invalid, normally the invalid group is automatically removed. However, due to how events are handled, there's still a query that is performed, with the invalid group. In Lokiless mode, this can even cause an error due to no matching metric found (although the root problem exists even with Loki enabled, it's less visible).
Steps to Reproduce:
1. Install netobserv in Lokiless mode 2. Go to Topology and set scope to Owner 3. In display options, set Group to Namespaces 4. Switch Scope to Node
Actual results:
Error is displayed because Prometheus cannot handle this query
Expected results:
No error: Prometheus should be able to handle it.
In fact, two queries are run: a first one that still has the Groups option, and a second one without the Groups because they were found to be invalid (Namespace groups on node scope is invalid). But the first query is still executed and this is where the error comes from. If we refresh the page, we'd see the error go away.
- links to
-
RHSA-2024:133664
Network Observability 1.6.1 for OpenShift
- mentioned on
