How Prometheus gets metrics data? is it Pulled by Prometheus or Promitor pushes it?

[testskills]

Apology for basic question. But I do see 2 configurations. Which exact configuration is responsible for getting scraped metrics data into Prometheus.

Prometheus:

  - job_name: promitor-agent-scraper
      metrics_path: /metrics
      static_configs:
        - targets:
          - promitor-agent-scraper.default.svc.cluster.local:8888

Promitor-scraper:

metricSinks:
      prometheusScrapingEndpoint:
        metricUnavailableValue: "NaN"
        enableMetricTimestamps: "true"
        baseUriPath: "/metrics"
        labels:
          transformation: "None"

[tomkerkhove]

Prometheus is in charge of pulling. The configuration in Promitor is simply to expose an endpoint that can be scraped, hence why the configuration is called prometheusScrapingEndpoint.

[testskills]

Thank you!

[testskills]

Just a follow up question. Where exactly Promitor stores the scraped data till the time Prometheus pulls it? And for some reason Prometheus is down then will there be any data loss.

[tomkerkhove]

Promitor stores it in memory and refreshes the data on the next pull, it is up to Prometheus to pull and store the data as it is designed for it.

If Prometheus is down, it is not possible to pull the info and thus it will not have the info.

Prometheus supports push gateways as well but that is not supported. We had it on the backlog but given there was no ask it has been closed.

Feel free to comment on it if this is a blocker, but it would require a concrete customer scenario that is committed to using Promitor.

[testskills]

Thanks! It’s not blocker . Curious to know how it works.

[tomkerkhove]

No worries and happy to help!