The docs for Docker show a snippit about ha for leader.
Cribl recommends deploying the Leader on stable, highly available infrastructure, because of its role in coordinating all Worker instances.
I do not see this on the Distributed Deployment docs page so I am curious if this should apply to all instance of a Leader no matter then deployment type. I feel it should but would like to know the best practices for Leader.
So digging into this more after no response from community I think that a Leader should be highly available no matter what the deployment strategy is.
Since your leader node is the configuration and monitoring platform for all of your workers/worker groups, then I would say it makes sense for it to be highly available. What’s cool, is that with GitOps and some VIP magic, you could probably implement some DR options in there too.
We don’t have an officially supported method for HA leader nodes. Yes, you can get creative and come up with your own solution as some customers have. FYI, true HA support is in the works but there is no ETA for it atm. Keep in mind that in the majority of use cases, a leader going down won’t result in worker nodes not processing data. You can fairly quickly an easily restore a leader’s configs from a remote repo as documented here.