Best HA solution for Promethues

[sreehari M V]

Hi All ,

Greetings,

Can you please suggest a best High Availability solution for Prometheus.

Server count: 750

OS : RHEL 7

Exporters: node_exporter, process_exporter and JMX exporter

Thanks and Regards,

Sreehari M V

[Ben Kochie]

I recommend Thanos[0]. It provides multi-server deduplication, scaling, and long-term storage.

[0]: https://thanos.io/

--
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/a647071f-a8ca-4afe-ba8d-5162589fe207n%40googlegroups.com.

[sreehari M V]

Thank you Ben for the information. Does this setup require SSD disks ?

Could you share installation steps in RHEL 7 if possible?

Thanks and Regards,

Vinod M V

[hark...@timescale.com]

You can try TimescaleDB for this. TimescaleDB's Promscale has HA configuration for handling Prometheus in HA mode.

[Ben Kochie]

Depends on a lot of factors. How many metrics, how many samples per second, how many datacenters, what kind of object storage you have, etc.

[Aliaksandr Valialkin]

Take a look also at VictoriaMetrics - see https://victoriametrics.github.io/#high-availability and https://victoriametrics.github.io/Cluster-VictoriaMetrics.html#high-availability 

--

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/a647071f-a8ca-4afe-ba8d-5162589fe207n%40googlegroups.com.

--

Best Regards,

Aliaksandr Valialkin, CTO VictoriaMetrics