10min_cpu_iowait but my CPU(s) are not showing anything that correlates to what Netdata warns me

Help
,
1

Hi there,

Environment

I’m running Netdata in a docker container and let it update-check on a daily basis:
docker pull netdata/netdata && docker stop netdata && docker rm netdata && docker run -d --name=netdata …
:warning: This might have been happened shortly after I got the warning below…!?

Problem/Question

Starting some days ago (Jan, 30th) I got this warning from one of my netdata-agents:

  • 10min_cpu_iowait with 40.63 %
  • some minute later: a clear messages/info-mail 10min_cpu_iowait with 16.59 %
  • some hours later: another warning 10min_cpu_iowait with 40.54 %

What I expected to happen

As I can see in my “System Overview” iowait is not significantly high, so that I’m still missing the finale clear messages/info-mail.

Why does it not clear this warning? And/or how can I clear this manually?

Kind regards
Thomas

2

Hi, @thomasmerz.

10min_cpu_iowait alarm has delay option, which affects sending notifications (adds a delay as you may guess):

delay: down 15m multiplier 1.5 max 1h

See the option description in our documentation.

And as we found out in your other topic: if you install Netdata in a container on macOS you monitor Linux VM, not the host.

3

In this case this is a “normal” Linux VM running a Netdata Agent in a Docker Container :wink:

Many thanks for this computational aid :slight_smile: Just that I have understood it correctly:
The warning will only be cleared, if there had been or will be a period of at least 22 minutes and 30 seconds (15 mins x 1,5) without a high iowait-timeframe again (which is currently not the case - so everything seems to be fine with Netdata and I have now a better understanding).