monitoring server node from F5 load balancer
I'm setting up Couchbase for our environment and have a question about monitoring the server nodes that I could not find in the documentation or the recent monitoring topic (http://www.couchbase.com/webinar/monitoring)...
- we have a cluster of 4 servers behind an F5 load balancer.
- auto-failover setting disabled.
- example URL we will tell our web app to use is like this: couchbase.domain.com.
- couchbase.domain.com will go to our load balancer and it will send via round-robin to 1 of the 4 nodes (servernode.domain.com:8091)
The F5 load balancer has a health check monitor. This ensures all nodes are up and healthy. Typically, the basic HTTP monitor will do a GET request for / and as long as it receives a 200 HTTP status, it is OK.
We need to setup an F5 health check for our Couchbase servers, but I don't know what to use. GET / does not accurately check the health of the active server nodes.
I was thinking of doing a GET /pools/default/buckets/bucket1 and look for word "healthy" or else, mark that node down. Any suggestions?