I received an email saying that one of my instances went out of memory. Looking at Grafana, it seems this isn’t the case, but there was an issue as there is a gap in the graphs.
Out of memory: Killed process 329 (query-server) total-vm:1129292kB, anon-rss:131176kB, file-rss:0kB, shmem-rss:0kB, UID:0 pgtables:360kB oom_score_adj:0
To give some contest, I was doing some stress test sending more than a hundred request per second.
Can someone give a hint of what could be the issue?
I would trust OOM over grafana. The gap in grafana is likely due to the machine crashing and the metrics being lost.
I was also getting those notifications regularly since few weeks. I’d recommend adding some swap to overcome RAM usage spikes:
It makes sense, thanks!
Thanks! I will take a look into it.
After some changes, I have done more tests, 150 requests per second, and the memory metrics are now flat.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.
