How to calculate disk resource will consume

[wangweihong]

For example , I scrape a target every 5 second and store data in a 40GB disk. I want to calculate how long will it spend to consume all my disk resource. What should I do?

[beorn7]

That also depends on the amount of samples you have per scrape. And on the compressibility of your sample data. Typical production data takes 3 to 4 bytes per sample on disk, plus some baseline space for the indices. You could also configure varbit encoding, see https://prometheus.io/blog/2016/05/08/when-to-use-varbit-chunks/ , which might compress down to 1.3 bytes/sample (at the price of query performance).

[wangweihong]

@beorn7 Thanks for your reply. I'm new to prometheus and don't know where to get the amount of samples. I set up prometheus to monitor itself and try to find some infos of sample , but I don't know what its metrics mean. Can you please tell me where to get the amount of samples and the docs about promethus's metrics?

[beorn7]

If you go to the /metrics endpoint of a prometheus server, you see all the metrics with doc strings.
The metric you are looking for is prometheus_local_storage_ingested_samples_total.

[elluvium]

I'm using openshift/prometheus:v2.2.1 image with local target and there is no such metric prometheus_local_storage_ingested_samples_total. What I'm doing wrong?

[beorn7]

You are using Prometheus 2.x, which is very different from Prometheus 1.x. Remember, this ticket was closed 2 years ago. I guess now you would use prometheus_tsdb_head_samples_appended_total.

It makes more sense to ask questions like this on the prometheus-users mailing list rather than in a GitHub issue. (A closed GitHub issue is even less suited.) On the mailing list, more people are available to potentially respond to your question, and the whole community can benefit from the answers provided.

[elluvium]

Ok, got it. Much appreciate for your answer even on closed ticket.