You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Every once in a while our junos exporter instance runs out of file descriptors and that looks like this. We have 14 devices configured currently, but we see the default ulimit of 1024 being exhausted with what I assume are stale sockets.
root@prometheus:/proc/1479/fd# systemctl status prometheus-junos-exporter.service
● prometheus-junos-exporter.service - Prometheus JunOS Exporter
Loaded: loaded (/etc/systemd/system/prometheus-junos-exporter.service; enabled; vendor preset: enabled)
Active: active (running) since Wed 2022-01-19 09:10:02 CET; 5 days ago
Main PID: 1479 (junos_exporter)
Tasks: 16 (limit: 19105)
Memory: 79.2M
CGroup: /system.slice/prometheus-junos-exporter.service
└─1479 /var/lib/prometheus-junos-exporter/go/bin/junos_exporter -config.file /etc/prometheus/junos-exporter.yml
Jan 24 15:44:43 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:43 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:44 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:44 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:45 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:45 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:46 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:46 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:47 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:47 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:48 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:48 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:49 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:49 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:50 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:50 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:51 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:51 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
Jan 24 15:44:52 prometheus.man-da.de junos_exporter[1479]: 2022/01/24 15:44:52 http: Accept error: accept tcp [::]:9326: accept4: too many open files; retrying in 1s
This seems to be strongly related to the failing ssh connections mentioned in #167, so I expect that failing ssh connections aren't properly being closed.
Every once in a while our junos exporter instance runs out of file descriptors and that looks like this. We have 14 devices configured currently, but we see the default ulimit of 1024 being exhausted with what I assume are stale sockets.
The text was updated successfully, but these errors were encountered: