A prometheus-style metric exporter that watches arbitrary processes
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2016-11-14 14:43:04 +00:00
.gitignore git ignore some detritus 2016-10-27 21:56:54 +01:00
cpanfile Commandline flags for port and config file 2016-10-27 22:03:11 +01:00
install-deps.pl cpanfile + snapshot of install-deps.pl 2016-10-27 21:55:44 +01:00
PERL-AS-ROOT.md Expand PERL-AS-ROOT instructions 2016-11-14 14:43:04 +00:00
process_exporter Take labels into account when registering a collector; allow multiple having the same name 2016-11-03 15:06:42 +00:00
README.md Add note to README about being unable to scrape process_open_fds of other users' processes when not root 2016-10-31 17:23:42 +00:00

process_exporter

Export Metrics About External Processes

This prometheus-style metric exporter exposes the basic process-wide metrics about a pre-configured set of other processes to watch.

This exporter does not require root privilege to run; however, there are are certain metrics it is unable to scrape about processes run by other users if it is not root. These are:

  • process_open_fds

Configuration

It is configured by a YAML file that gives names and ways to find external processes, most likely by reading a pidfile.

processes:
  my-server:
    pidfile: "/var/run/my-server.pid"

Exported Metrics

Process-wide metrics about the monitoring processes are exported under the exported_process name prefix, using a label called process_name whose value is the name given in the configuration file.

exported_process_cpu_seconds_total{process_name="my-server"} 1.234

See Also

There are other process-metric exporters by the same name with different behaviours.

This implementation differs from the above primarily because it has a pre-determined set of processes it will watch, configured by the YAML config file. These other exporters work by scraping all the processes found, looking for patterns of behaviour that make them interesting enough to export.