We are using EKS and one of our dedicated worker node for TimescaleDB (TSDB) went down. The node soon displayed unreachable taints and a new node got scheduled in place of it soon after.
We want to investigate further as to why the node went down. We have an idea that the TSDB pod was operating on high memory moments before the crash but we would like to be concretely point out the reason for the issue in order to fault-proof it for future.
Can someone suggest a direction to take? We already looked at the logs for the instance that went down, there was nothing that concretely points to a crash or node being unavailable.
AWS OFICIALAtualizada há 2 anos