There seems to be no malware, we ran chrootkit and rkhunter and there are no rootkits, no malware.
As an information, the server is not a homelab, it is a Contabo VPS, and it only hosts a Discourse container and a Prometheus/Grafana container.
The situation has gotten even worse since yesterday, though.
(12h graph)
Is there any kind of “low intensity” setup wrt stat recalculation?
I’m still looking at Sidekiq and I’m not seeing any specific pattern in the job names, it seems any process is at a risk of taking >2 seconds.
I’m starting to think looking at the scheduler isn’t very helpful.