[Bug]: Netdata stops collecting apache/nginx logs due to logrotate about netdata

[ilyam8]

Hey, @bgabler. Unfortunately, I can't reproduce the issue.

I was surprised by this issue as I was confident that the log rotation handling was working - there is even a test case for it.

Anyway I decided to do a manual test and had no problem:

• install nginx.
• restart netdata and make sure it tails /var/log/nginx/access.log.
• run sudo logrotate --force /etc/logrotate.d/nginx.
• check web log charts.

---

I think there is confusion, because

template: NGINX
      on: nginx.requests

has nothing to do with logs/rotation. The nginx.requests metrics collected by go.d/nginx.

It sends HTTP requests to the NGINX location stub-status, which is a built-in location that provides metrics about the NGINX server.

[bgabler]

It's so strange, as soon as we restart netdata it starts collecting again.

I modified our go.d file to have:

autodetection_retry: 3

This was commented out before.

[ilyam8]

autodetection_retry is

Recheck interval in seconds. Zero means no recheck will be scheduled.

Before starting a data collection job, the plugin checks whether the data can be collected and does not run the job if the check fails. So I see no how this option will help you.

Consider checking logs, always a good idea when debugging a problem.

• tail Netdata logs

journalctl --namespace=netdata -f

• force nginx logrotate
• see logs