Post 903341 by wheeler@linuxrocks.online
(DIR) More posts by wheeler@linuxrocks.online
(DIR) Post #896261 by fribbledom@mastodon.social
2018-11-01T05:45:39Z
0 likes, 0 repeats
How do you monitor system(d) services on your Linux server?Bonus points for being able to get notified about errors via mail _and_ IRC.
(DIR) Post #896283 by chaomodus@radical.town
2018-11-01T05:47:14Z
0 likes, 0 repeats
@fribbledom Well icinga2 is what I'd do, and irc and email are easish, but it'st not like off teh shelf ready to go, it'd be a little doing (fun doing but it depends on what you're into)
(DIR) Post #896370 by codythedragonrude@mastodon.social
2018-11-01T05:59:06Z
0 likes, 0 repeats
@fribbledom you don't monitor systemDeath it monitors you 🙃
(DIR) Post #896393 by ben@mastodon.lubar.me
2018-11-01T06:01:48Z
0 likes, 0 repeats
@fribbledom I monitor my Linux server by getting a 52X error message from Cloudflare whenever it mysteriously dies and I try to check Mastodon.
(DIR) Post #896415 by rus@mastodon.technology
2018-11-01T06:03:33Z
0 likes, 0 repeats
@fribbledom doesnt systemd itself have internal monitoring that will send out emails on failure? https://serverfault.com/questions/694818/get-notification-when-systemd-monitored-service-enters-failed-statebut theres also monit, nagios, munin, prometheus, graphanai cant seem to find a modern replacement for nagios (ive moved from munin to prom/graf) but nagios certainly does email & irc notificationsit all depends on what youre monitoring and why really but right this second if i wanted email & irc notifications of services running on any kind of server id go nagios
(DIR) Post #896519 by flussence@nulled.red
2018-11-01T06:15:21Z
0 likes, 0 repeats
@fribbledom usually I'll notice straight away when something's wrong because my entire server is a fragile house of cards and any error causes my mail client to yell at me that it can't connect
(DIR) Post #897193 by dlek@x0r.be
2018-11-01T07:19:33Z
0 likes, 0 repeats
@fribbledom Syslog everything. It's a well-supported transport that's been around forever. syslog-ng and rsyslog can both many needs.Log the things you don't understand, and for the rest, ignore or create an actionable trouble ticket. If emergent, page someone.Tie related events together and manage with vigilance. This includes daily perusal of the unclassified logs to determine appropriate action.Syslog-ng provides a pattern database of many known logs.
(DIR) Post #897205 by dlek@x0r.be
2018-11-01T07:20:13Z
0 likes, 0 repeats
@fribbledom Syslog everything. It's a well-supported transport that's been around forever. syslog-ng and rsyslog both meet most needs.Log the things you don't understand, and for the rest, ignore or create an actionable trouble ticket. If emergent, page someone.Tie related events together and manage with vigilance. This includes daily perusal of the unclassified logs to determine appropriate action.Syslog-ng provides a pattern database of many known logs.
(DIR) Post #897239 by thunfisch@chaos.social
2018-11-01T07:22:14Z
0 likes, 0 repeats
@fribbledom i run a central zabbix server that collects all sorts of stats on the increasing army of servers i administrate.you can do checks on services using userparameters in zabbix agent.alternatively: systemd has its own options to send mails when something goes fubar.
(DIR) Post #897757 by rodti@mastodon.scot
2018-11-01T08:16:23Z
0 likes, 0 repeats
@fribbledom I use monit with email alerts to my Zammad support software. Zammad's monit integration opens a ticket when a service is down, and closes it again if it comes back up. Also monit provides a (passworded) status/uptime page on its own web server.
(DIR) Post #897987 by cm@chaos.social
2018-11-01T08:38:29Z
0 likes, 0 repeats
@fribbledom I use check_mk, which is an agent/configuration thing for nagios and makes managing nagios so much easier. should also work with icinga.
(DIR) Post #903341 by wheeler@linuxrocks.online
2018-11-01T15:13:57Z
0 likes, 0 repeats
@fribbledomNetDatarsyslog + GraylogPrometheus + GrafanaCheck_MK
(DIR) Post #904505 by lzap@mastodon.social
2018-11-01T16:24:01Z
0 likes, 0 repeats
@fribbledom pcp.io by SGI HPC. Maintaned by Red Hat and the community. Used by Netflix and others. It's pretty unique, high-resolution, highly flexible.