Increase runner pod memory usage from version 12.2 to 12.6 (Kubernetes Executor)
This bug was reported by a customer (internal link). This issue will be updated with more screenshots/monitoring metrics.
After performing direct upgrade of the runner (performed on 01-20-2019) with a rebase of 12.2 to 12.6, including this patch. An increase in general usage and spikes have been seen for memory as high as 2GB for ideal runners (with a limit of 1Gi).
- Is this memory increase expected for newer version of the runner using the Kubernetes Executor?
- Is it possible for a container to not fail when reaching memory above it's current limit?
Steps to reproduce
Upgrade the runner from 12.2 to 12.6 with this patch
Increase memory usage in idle and working runners
No increase in memory usage in idle and working runners
Relevant logs and/or screenshots
Kubernetes Executor(Helm Chart) GitLab 12.6.4-ee
Add your configuration here
Used GitLab Runner version
12.6.0 (with patch )