I will start implementing a kind of back pressure feature in the near future.
The reason for doing this is that in error scenarioes where a couchbase node is done or not responding it can cause my entire application to become unresponsive. My current debugging efforts shows that it is caused by excessive garbage collection which is happening because a lot of operations has queued up and requests just keeps coming from the load balancer.
Would you be interested in adding this feature to the driver as a kind of optional feature. I would off course maintain it for the future or do you think it is a kind of feature creep?
I will start implementing a kind of back pressure feature in the near future.
The reason for doing this is that in error scenarioes where a couchbase node is done or not responding it can cause my entire application to become unresponsive. My current debugging efforts shows that it is caused by excessive garbage collection which is happening because a lot of operations has queued up and requests just keeps coming from the load balancer.
Would you be interested in adding this feature to the driver as a kind of optional feature. I would off course maintain it for the future or do you think it is a kind of feature creep?