Extend Prometheus exporter with extra / more data points
We want to enable better integration with 3rd party dashboard tooling, APM integrators and other by providing more data points on the Checkly Prometheus exporter. Prime candidates are:
connection phase metrics for API checks, e.g. TTFB, dns
more metrics for browser checks like web vitals, TTFB etc.
Log in to comment and vote
Comments5
BaCaRoZzo
Sep 14, 2023
Thank you very much for these improvements!
Our company rely heavily on
prometheusandAlertmanagerfor alerting so adding more metrics or just more detailed metrics is very welcome.Here are some suggestions I would like to give:
Can you add the
frequencyof the check as a label? As a matter of fact we useforclauses inprometheusto let a check fail a predefined number of times. If we can export thefrequencyalong with the other information we can use that to create different rules with differentforclause without having to filter on anything elseThe overall
checkly_check_statusis very useful but sometimes we can be interested in having the status for a single location. Could you add a separatecheckly_check_statusfor each location?checkly_check_resultis very detailed and very much needed. At the same time in this case having an overall/average metric can also be useful. Although we don't have a specific use case in mind and it can be easily generated viaprometheusconstructs. Still nice to avoidavg/avg_over_time, if the checkly exporter provides it out of the boxSometimes it would interesting to export also some information from the payload for further analysis down the line. For instance we have a simple
pythonscript whichGETs from an endpoint and returns a metric containing a label equal to a field of the payload. Checkly allows to define a predicate over a field but there is no way to specify that the field must be exported as a label toprometheus. I understand this proposal can become pretty complex, pretty soon. Still, I thought it would be a nice food for thoughts.