What is the optimal value one should set for RequestTimeout parameter for a busy website for failover to occur quickly in cluster environment?
Are there any other parameters which will help agent to failover to next policy server cluster?
Any other parameters which can be tune for busy site?
There is no optimal standard value for RequesTimeOut as it depends on many factors. '
If you are using cluster configuration for Policy servers, you can configure failover threshold (percentage) to indicate when the cluster failover should happen.
The recommended way is to add multiple policy servers in the cluster and configure lod balancing to distribute load.
We do have cluster configuration and for each cluster we have min 5 policy server in mix with the threshold % set to 60%
When the cluster failover occur, we notice that web agent halt all the request for 60sec. That's where we are looking to see if RequestTimeout parameter can be adjusted to avoid that delay or any other parameter which will help agent to failover immediately and not causing all the request to Q up.