At the moment, workload progess is governed by the cache configuration parameters of usage trigger, dirty trigger and updates trigger.

When cache reach any of the configured trigger levels, backpressure is applied to the workload by engaging application threads for eviction. In some cases this behavior has the affect of workload stalls.

Allowing workload to progress when cache reaches configured trigger levels, will result in more memory usage, could lead to more memory fragmentation. Also, as the dirty content in the cache will increase, the checkpoint times will increase and in some cases could result in increased recovery times, directly impacting the restart duration.

This ticket is to investigate the impact of making cache tolerant to configured trigger levels.