As described in this old post , we have two VMs hosting an Event Server service. One is idle and the other one is effectively receiving the audit logs from our deployed APIs.
If the "hot" one crashes (big crash), we would need to manually change the infrastructure configuration in the Deployer, pointing the audit logs configuration to the IP address of the other VM and then redeploy all our services (pain point shared in this other old post).
As soon as we are "well trained" internally, these manual steps wouldn't be too much complicated and everything would be fine quickly, but we would still have some suboptimal aspects: 1) Manual steps required, 2) Humans might be not available, 3) Some audit logs would be lost in the meantime.
The suggestion here is for the Deployer infrastructure configuration to have a primary config and an optional secondary config (a different IP address where another Event Server is running). This would pretty much solve our problem, but maybe there is a smarter way of achieving the same goal for multiple Dataiku clients.