cloud-pipeline
cloud-pipeline copied to clipboard
Issue #2568 Double-check active VMs for missing node notification
trafficstars
This PR is related to the bug described in the comment https://github.com/epam/cloud-pipeline/issues/2568#issuecomment-1091858826
Cloud-Pipeline node down operation consists of 2 steps:
- removing k8s node;
- terminating VM in the cloud.
It might lead to false-positive missing node notifications in case monitoring cycle started during that transition.
PR brings additional fetching of active VMs from the cloud provider at the end of checking iteration - all the notifications on the missing nodes for VMs, that are already terminated, are dropped.