Hey team, I am using cockpit on a whole bunch of servers to the point that it’s a bit of a chore to select each server manually in cockpit just to check e.g. disk space / mem usage. On all other distros there is a cockpit-dashboard package (Dashboard tab is missing. · Issue #6486 · cockpit-project/cockpit · GitHub) that allows one-page view of all servers but I wasn’t able to find instructions on how to set this up on Rocky. The link / github discussion suggests it’s doable but I can’t work out how. Many thanks in advance.
The dashboard component of Cockpit was removed in 234. Unless upstream cockpit devs bring it back or there are users who wish to maintain it in say EPEL or otherwise, it won’t be available in Rocky Linux or other distributions.
If other distributions are still providing this, it is likely they are maintaining it themselves.
This is a common problem with monitoring tools, you need a way to separate the collector from the monitor, e.g. performance co-pilot (the RHEL way of doing it).
Many thanks - I eventually found this too. Was the most useful feature, I am not sure the whole thing is worth using without it.
Thanks gerry666uk I’ll check out the performance co-pilot.
But it raises an interesting question; what exact mechanism was the cockpit dashboard using to monitor the servers? They could have used performance co-pilot and integrated it with cockpit, but if they did, why are they now removing it?
For me, something put me off using cockpit; anything with a nag screen makes me suspicious, and I didn’t like the look of it for virtual machines.
Probably because it’s more of a server management solution, rather than a monitoring solution. You get by far better metrics using a proper monitoring system like CheckMK, Zabbix or whatever, with historical views going back X amount of days/months/years.
I assume you mean “cockpit” here, as opposed to performance co-pilot?
So yes, maybe they decided it should not have monitoring anymore.
Regarding Zabbix and friends, and retrospective anaysis, I think RH would say that performance co-pilot offers this too.
It applies to cockpit and performance co-pilot. That is not a monitoring solution. Monitoring solutions are far more complex, provide notifications, etc, etc. I suggest you compare to see how vast the difference is. Cockpit is a server management/administration tool.
So the word ‘and’ that is in bold, is saying performance co-pilot is a server management solution and not a monitoring solution?
Take a look at: Checkmk features enable you a powerful IT monitoring and Zabbix features overview and you will see why cockpit or co-pilot is not a monitoring solution. Basic CPU/memory/disk is not monitoring - especially since it doesn’t send email/sms or other types of alerts without you actually looking at the server - at which point it’s too late. If I get notifications that my server is slowly running out of disk space before it happens, I can actually do something about it. And I don’t need to keep looking at it every day, since the monitoring solution does it for me, and I can concentrate on far more important things.
CheckMK Zabbix or others monitor and provide far more detail, not just for CPU, ram, disk, but also processes, services like Apache, Nginx, MariaDB, Postgresql, etc, etc, etc. And I can see exactly which process is using too much ram and running out of control. You don’t see that with cockpit.
If I log into cockpit and my server has already ran out of disk space, then it’s already too late.