On Tue, Oct 21, 2025 at 1:42 PM G.W. Haywood <[email protected]>
wrote:

> It doesn't look like there will be a lot of work involved, but there
> seem to be conflicts in the suggestions which may need to be resolved.
> I've never met Prometheus, but I do use Icinga extensively so I guess
> it won't be terribly unfamiliar.  Can anyone here help me with this?
>

Prometheus and Incus have very little in common. :)  Prometheus has more in
common with something like MRTG (in that it implements the polling and data
storage components).

Prometheus records data in a time series database, and provides a query
language for retrieving that data in ways that make it easily graphable and
easy to generate alerts from.   It's commonly paired with Graphite and
various other tools that feed into push notification (alerting) software.

Prometheus operates by having an "exporter" run on remote systems that
provides an HTTP(s) endpoint that dumps current statistics counters
(key/value).  Exporters are not dissimilar to SNMPD, just with a different
interface and security profile.   Prometheus scrapes the data from whatever
exporters it's configured to access, at regular intervals, and stores the
retrieved statistics in its time series database to be looked at later by
some other process (e.g. Graphite).

Some software comes with an exporter built in (for full circularity,
Graphite provides a Prometheus exporter endpoint, and then provides its own
Dashboard showing its own performance based on the data stored in
Prometheus).  For other software there may be an additional process that
collects and collates data into statistics and provides the endpoint (e.g.,
see the "Node Exporter", which provides system statistics for an OS... this
is really the only place there is overlap with Icinga/Nagios).
BackupPC_Admin?action=metrics?format=prometheus
implements an exporter for Backup PC.


I have only done a very brief review of the issues you mention... It looks
like the conflicts come from different people trying to solve the same
ongoing issue in different ways, depending on their personal needs.  My
personal opinion is that "fix this problem by fixing the value" is a better
approach than "fix this problem by removing the value."   Beyond that any
conflicts arising in what is considered "fixing" a broken statistic is
probably going to require some more knowledge of internals than I have time
to dig into right now.
_______________________________________________
BackupPC-users mailing list
[email protected]
List:    https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki:    https://github.com/backuppc/backuppc/wiki
Project: https://backuppc.github.io/backuppc/

Reply via email to