Bug: compute node's logs have two sets of mismatching timestamps

neondatabase / autoscaling

Postgres vertical autoscaling in k8s

Apache License 2.0

161 stars 21 forks source link

Bug: compute node's logs have two sets of mismatching timestamps #740

Open Omrigan opened 9 months ago

Omrigan commented 9 months ago

Environment

Production

Steps to reproduce

Open logs and find a line with ts=.., e. g.

2024-01-16 00:01:32.140 ts=2024-01-16T00:01:28.288Z caller=proc.go:250 msg="Excluded databases" databases=[]

Expected result

The timestamps should match

Actual result

They don't.

Other logs, links

This either happens due to imperfect clock sync, so should be fixed by #732 or with slow log collection, so should be fixed by #724

sharnoff commented 8 months ago

@Omrigan seems like this is handled by other issues, should this be closed?

Omrigan commented 7 months ago

Still the case:

2024-03-27 08:57:38.394 
ts=2024-03-27T08:59:10.459Z caller=klog.go:124 level=error func=Errorf msg="Error gathering metrics: [from Gatherer #1] [collector=neon_collector,query=lfc_misses] pq: relation \"neon.neon_lfc_stats\" does not exist"