Monitor github actions runner with telegraf + prometheus

[Mic92]

The github actions runner API can be used to external monitor
if your github actions runner are online.
The endpoint for an organisation is https://api.github.com/orgs/<ORG>/actions/runners,
where ORG is the name of the origanisation.
Accessing the API requires github token of a github account that has admin access in the organisation (admin:org scope).
It can be generated here: https://github.com/settings/tokens

The following telegraf configuration will periodically poll this API endpoint:

[inputs.http]
# file that contains the github token generated above (starts with `gph_`)
bearer_token = "/var/lib/telegraf/github-token"
data_format = "json"
fielddrop = ["labels_*", "id"]
json_query = "runners"
json_string_fields = ["status", "busy"]
tag_keys = ["os", "name"]
# Replace <ORG> here with the actual github origanisation
urls = ["https://api.github.com/orgs/<ORG>/actions/runners"]

From telegraf it can be exported to prometheus or influxdb.

The following example alert condition in prometheus checks if there at least two
online github runner: count(http_busy{status="online"}) >= 2

Cheers!

[Mic92]

I also recently learned about https://github.com/Spendesk/github-actions-exporter

[callum-tait-pbx]

#1116 probably worth linking the existing observability issue here too as it is loosely related to this wiki and may down the road have related comments