Hi All,
Currently I am working with a Prometheus -> TimescaleDB stack for long term storage. We make use of a few custom down-time metrics for SLA reporting.
For long term storage, I am not interested in keeping all metrics collected (eg; node_exporter), just the custom metrics.
I do not see a way to retain on a per-metric basis but thought I would confirm. At the moment we just have an SQL job that removes everything from the database aside from the custom metrics, but this doesn't seem optimal.
Is there a way to perhaps configure either of the following?
1) Only push certain metrics to TimescaleDB, so only push the custom metrics to the remote end-point.
2) Custom retention periods per metric, so discard select metrics of no interest after X time.
Cheers,
Ruairi