We’re adding a Couchbase health check to our application that attempts to connect to Couchbase and retrieve a document by key.
I’m testing some failure scenarios. One of these is to change my cluster pool URLs to invalid URLs and then running the healthcheck. Despite changing various timeout settings, the call to OpenBucket is taking over 90 seconds to run. Ideally I want this to timeout within a few seconds so that we can be alerted to the health of the application asap.
What is the correct setting to use to reduce the OpenBucket timeout?
There is not an OpenBucket timeout per se’; opening a bucket requires bootstrapping and depending upon your cluster size and the #of connections in your pool varies considerably. Also, the bucket type itself may haven an impact on the time to bootstrap.
Note that the client will attempt every invalid url using two different (HTTP streaming and CCCP) protocols to connect - that is probably why its taking 90 seconds. The client agressively trys to connect because that is the most common scenario requested; people want there apps running.