systemd-units-failed: Add option to include failed unit names in first output line

[metrax]

Describe the solution you'd like

Currently, the plugin /usr/lib64/nagios/plugins/systemd-units-failed only reports a generic message in the first output line, e.g.:

There is 1 failed unit.

When monitoring multiple systems, this message is not very helpful, as it requires opening the plugin’s extended output or performance data to see which service actually failed.

I propose adding an optional parameter, for example --include-unit-names (or similar), that changes the first output line to include the failed unit names directly when there are only a few of them.

Example desired output:

1 failed unit: letsencrypt-wrapper.service

or if multiple:

2 failed units: letsencrypt-wrapper.service, postfix.service

This would make it easier to immediately identify failed units from dashboards or SMS alerts without needing to expand the plugin details.

Additional context

The feature should be optional to maintain backward compatibility with existing parsing or alerting setups.
Example command for new behavior:

/usr/lib64/nagios/plugins/systemd-units-failed --include-unit-names

[markuslf]

That's a sensible suggestion, and I don't think that there is a need for backwards compatibility in a plugins' message output.

[metrax]

That's a sensible suggestion, and I don't think that there is a need for backwards compatibility in a plugins' message output.

Thanks for the quick feedback, much appreciated!
Agreed — no strong need for backward compatibility here.
The shorter “1 failed unit: ” format would be a great improvement for alert readability, especially in dashboards and SMS alerts.
Looking forward to seeing it implemented!

[markuslf]

Implemented in d669cbf. The first output line now shows the failed unit names directly, for example 2 failed units: openipmi.service, smartd.service. No extra parameter needed, the new format applies by default.