From 8db7d12cc5ccea800f0d357c369206b8c276b706 Mon Sep 17 00:00:00 2001 From: SuperQ Date: Fri, 21 Nov 2025 08:25:42 +0100 Subject: [PATCH] Improve HA FAQ entry Update the high availaibliy FAQ entry with information about data deduplication. Remove wording about memberlist as it's confusing to readers that it's part of the Prometheus HA design. Signed-off-by: SuperQ --- docs/introduction/faq.md | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/docs/introduction/faq.md b/docs/introduction/faq.md index fa9c5f168..ee7e106eb 100644 --- a/docs/introduction/faq.md +++ b/docs/introduction/faq.md @@ -33,9 +33,12 @@ There are always trade-offs to make when running services, and Prometheus values ### Can Prometheus be made highly available? Yes, run identical Prometheus servers on two or more separate machines. + +Systems like [Thanos](https://thanos.io) can deduplicate the data. Many of the [external systems](/docs/operating/integrations/#remote-endpoints-and-storage) that support [remote write](/docs/prometheus/latest/configuration/configuration/#remote_write) can also deduplicate the data. + Identical alerts will be deduplicated by the [Alertmanager](https://github.com/prometheus/alertmanager). -Alertmanager supports [high availability](https://github.com/prometheus/alertmanager#high-availability) by interconnecting multiple Alertmanager instances to build an Alertmanager cluster. Instances of a cluster communicate using a gossip protocol managed via [HashiCorp's Memberlist](https://github.com/hashicorp/memberlist) library. +Alertmanager supports [high availability](https://github.com/prometheus/alertmanager#high-availability) by interconnecting multiple Alertmanager instances to build an Alertmanager cluster. ### I was told Prometheus “doesn't scale”.