My cluster made of 2 CASes has recorded a maximum 12.1M sessions at some point this week. My question is; do we exceed the capacity of the cluster (2*6M sessions)?
Does the capacity report shows the total number of sessions across all CASes?
Running select count(*) from rtmsession does report a value close to 12M on each CAS.
Solved! Go to Solution.
The 6 M sessions criteria is obsolete. There're few other KPIs that determine the CAS capacity utilization in your cluster. This was a formula used in past to scale CAS servers, but not it is of less importance as one may have a server with 20 M of sessions run fast and some others may have system with less sessions but choking.
Most important KPIs on CAS are:
* data processing tempo - avg. data processing per reporting interval (1 or 5 minute) must stay around half of the reporting interval (optimal), but no more than 3/4 of the reporting interval (to avoid processing delay
* memory utilization in green on Diagnostics -> CAS Memory and restarts (to avoid CAS running garbage collection cycles often - that suspends data processing)
* no timeouts during the nightly tasks execution (DB maintenance, trends/baselines/avgs generation and so on)
* Your general perception if the DMI reports in your cluster provide decent response time or not (that depends on the report definition per se, but the real-time data for today should work pretty fast)
Thank you! The primary CAS free process memory is at about 13% and the web service calls are reported slow by an application team. Others indicators are within acceptable ranges. The database size has grown fast recently.