Skip to content

Reinstate 60s timeout in Cluster Prometheus

Matthias Käppler requested to merge 232786-reinstate-60s-timeout into master

What does this MR do?

Refs #232786

This does the same thing as !39318 (merged), but I missed a case in that MR, which is the Applications::Cluster::Prometheus wrapper which holds its own PrometheusClient instance and creates long running DeploymentQuerys.

In short, we are in the process of rolling out a feature flag (see linked issue ref) that will set stricter timeouts for any outbound HTTP requests. However, some parts of the application send requests that time out on the new defaults. I am therefore reinstating the previous 60s timeout explicitly in those to unblock the feature toggle. Eventually we should fix the root causes, which are the underlying slow queries to Prometheus.

Does this MR meet the acceptance criteria?

Conformity

Availability and Testing

Edited by Matthias Käppler

Merge request reports

Loading