[FR] Timeout before die

[mattthhh]

As the exporter is started in the Rsyslog config, when Rsyslog die for a specified reason (oom, privileges changes on some files/dirs, systemd crash), the exporter also die with it. It could happen that the reason is not explained by the previous stats exported, because the crash happens really fast and prometheus did not scrape the interresting stats from the exporter before rsyslog died.

I was wondering, with the help of this option in omprog, if the binary could get an option to add a timeout before he terminates (as --timeout N) to keep the HTTP server running and exporting the latest stats before Rsyslog died ?

Let me know if you found a better option than that.