Hello We are currently in the system design phase ...
# fleet
g
Hello We are currently in the system design phase for implementing osq and fleet for our server estate (~2000 servers) and we were asked to check if we need 1 fleet server or create a cluster to ensure high availability. Request if someone can help with this please. Is it a good idea to have this HA set up and if we should, how should the set up be (failover details, active passive set up etc.)
b
How are you deploying? Public cloud or your own hardware? Generally putting a LB/Reverse proxy like nginx or HAProxy to terminate TLS and forward onto the fleet severs is the best setup. If you’re running in AWS we have terraform code that will stand up a production ready setup out of the box.
g
we have our own data centre where the servers are hosted
b
yeah then I'd recommend using nginx or haproxy and a couple of servers running fleet (mostly for HA) -- for the record a single host running the fleet server could handle 2000 hosts without any issues