spiked cpu for ~20 mins, rebooted

i wonder if its just a lot of americans posting garbage at this time...

Show thread

@subleq What about the hole in the server stats two hours earlier? Was that another CPU spike that impacted monitoring?

@wasamasa ah yeah same thing, but wasn't as bad and the server somewhat recovered from it.

i'll have a looksie on better mitigation steps, like removing the overlayfs...

@subleq @wasamasa Was that the cause, though? SUnreclaim wasn't that high this time and the graphs don't suggest a memory problem.

Considering the high disk I/O and lack of swap usage, I'm more inclined to believe that the block storage had just stalled for a while.

@subleq @wasamasa Actually, maybe not. Peaked at 250 read ops/s and 14 MB/s during the downtime, which is pretty much the max of that block storage.

Wonder if that was a Masto or ES cleanup process or something.

@ayo @wasamasa ah

i did wonder if your backups were running

but i didnt check when i rebooted it

Sign in to participate in the conversation

A lonely little town in the wider world of the fediverse.