[5] Stability: Why does the pods get issues being terminated during node draining?

[robinelastisys]

What should be investigated. As a continuation of https://github.com/elastisys/compliantkubernetes-apps/issues/856, we now want to figure out why some pods get issues being terminated while nodes get drained on a deeper level:

harbor harbor-harbor-redis-0 0/1 Terminating 0 35h 10.233.118.108 pg-dev-sc-worker-2

influxdb-prometheus influxdb-metrics-retention-cronjob-wc-1636979700-mls7r 1/1 Terminating 0 46h 10.233.118.191 pg-dev-sc-worker-2

monitoring scan-vulnerability report-68cc76c95c-mn8jj 0/1 Terminating 0 37h 10.233.118.183 pg-dev-sc-worker-2

What technology/product/concept should be investigated? Node draining in general. How do we reduce the number of pods having issues being terminated?

What artifacts should this produce. An investigation doc detailing: What are the main reasons pods are having issues being terminated? How can we solve it?