DR set up for prometheus

Anoop Mohan

unread,

Sep 30, 2020, 5:41:36 AM9/30/20

to Prometheus Users

Hi,

Can anyone suggest what is the best method to implement disaster recovery for prometheus?

Our current implementation is east and west region prometheus are federating whole data with each other. So that we will have the other region data when switching to a region during DR. However, I'm not sure this is a recommended way for Prometheus DR, since the federation will be happening for a large amount of time series . Please advise if there are any alternate methods to achieve this?

Thanks & Regards,

Anoop

Kishor Bammidi

unread,

Sep 30, 2020, 5:50:16 AM9/30/20

to Anoop Mohan, Prometheus Users

Hi Anoop,

I have implemented Prometheus + Thanos with s3 for long term storage. We are running it on EKS in AWS though.

Regards,

Kishor.

--
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 prometheus-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/CAEQqSpexVn8U6L-aG4--nbZkriNNf0SwxE-%2BHtjm1cfAzxP_YA%40mail.gmail.com.

Brian Candler

unread,

Sep 30, 2020, 6:38:51 AM9/30/20

to Prometheus Users

What I use is multiple prometheus servers doing remote_write to a single instance of VictoriaMetrics. This gives me a global view, and long-term storage, and is really easy to setup. Grafana can query VictoriaMetrics in the same way as it queries Prometheus.

Reply all

Reply to author

Forward