Hello, I have a special use case and I want to ask the community opinion on the implementation
*Usecase*: I want to expose all metrics from a given cluster (not k8s cluster) to a prometheus server through only one endpoint. multiple components exposing metrics ---> single-endpoint --> main prometheus I have thought of 3 solutions but with some caveats *Solution1*: use a tool that merge metrics from multiple endpoints and expose them at once while injecting missing labels like instance and job *Solution2*: Setup a local prometheus server that scrape all metrics and use its /federate endpoint as a single endpoint for the main prometheus server to scrape *Solution3*: Use remote_read config option and point it to the local-prometheus/api/v1/read to ingest the local prometheus metrics *Caveats* *Solution1* is more like a dirty hack because metrics with different scrape intervals will be complicated to handle through only one endpoint *Solution2*: the main prometheus server will timestamp the metrics again from the federation (not sure if federate endpoint exposes metrics timestamp in that case i may honor timestamps) *Solution3*: This looks the best solution so far as I read that remote_read will be more efficient that scraping ? (I may be mistaken) Which solution you think it is the best, Any other suggestion is welcome Best regards -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/3e32f579-2772-4d30-9412-b033dcb149e9%40googlegroups.com.
