Limits to etcd cluster size

[arjunsingriuber]

Have there been any tests around how many hosts etcd can replicate to using RAFT?

[jmhbnz]

Hi @arjunsingriuber - Thanks for your question, please refer to our FAQ: https://etcd.io/docs/v3.5/faq/#what-is-maximum-cluster-size

Theoretically, there is no hard limit. However, an etcd cluster probably should have no more than seven nodes. Google Chubby lock service, similar to etcd and widely deployed within Google for many years, suggests running five nodes. A 5-member etcd cluster can tolerate two member failures, which is enough in most cases. Although larger clusters provide better fault tolerance, the write performance suffers because data must be replicated across more machines.

[arjunsingriuber]

Is there a limit on the number of observers?

[jmhbnz]

To clarify are you referring to a cluster learner member? Or etcd watch functionality?

[arjunsingriuber]

Sorry for the confusion. I am talking about etcd watch functionality.

[jmhbnz]

Thanks for the clarification, etcd is well known for the efficiency of it's watch capabilities, refer https://etcd.io/docs/v3.5/benchmarks/etcd-3-watch-memory-benchmark noting that those documented benchmarks are now getting old.

This is something you can perform basic benchmarks in your own environment with https://github.com/etcd-io/etcd/tree/main/tools/benchmark. I would encourage you to perform validation in your environment to properly understand the performance characteristics in your environment.

[arjunsingriuber]

Thank you