Lately we have detected that marin3r DiscoveryService pod was consuming lots of mmeory, even up to 22GB in production, producing to bring down openshift worker node where it runs, affecting to all pods from the node:

On lower environments (staging, dev, int...) we detected the same incremental memory usage with no stop, but lower numbers due to the size of lower environments.

By now we have set memory limits to this pod (we didnt had any cpu/memory req/limits auntil now).