Closed Porcupine96 closed 5 years ago
I have the same issue, somebody, please comment about it.
Same here, this seems to be a fairly important problem. Any solutions?
same problem. It seems it never closes the socket when it checks.
I found the issue, see below :)
I'm seeing this issue again using the per-compiled binarys (cachet_monitor_linux_amd64 and cachet_monitor_linux_386) with v3.0.
Ran this on two separate systems: Linux 4.14.67-grsec-grsec+ #4 SMP Thu Sep 6 01:23:09 PDT 2018 x86_64 x86_64 x86_64 GNU/Linux and Linux 2.6.32-954.3.5.lve1.4.59.el6.x86_64 #1 SMP Thu Dec 6 05:11:00 EST 2018 x86_64 x86_64 x86_64 GNU/Linux
Same issue here
time="2019-09-05T18:36:23Z" level=warning msg="Could not log metric! ID: 2, err: Post https://example.com/api/v1/metrics/2/points: dial tcp: lookup example.com on: dial udp: socket: too many open files"
Me too
Oh I see, the binary that has this issue fixed wasn't released......
Can anyone help me build the cachet-monitor binary to include the latest commits in master to fix this issue?
I added to the crontab a restart of the service each night, that fixed the issue. It's not the best solution ever, but given the time I had, it was a working and effective solution.
I've been restarting every hour with a crontab because of the php memory limitation of the shared host I'm using.
I'm having an issue - the cachet-monitor seems to leave the sockets opened.
At the beginning everything works fine but the number of used file descriptors constantly increases till it reaches the soft limit (1024) then I'm getting an error:
... level=warning msg="Could not log metric! ... socket: too many open files
Thank you for your time,