Sometimes the number of active requests exceeds throttle limits. This can happen for several reasons:
- A request uses too many resources, and TASM takes the exception action to move it to another workload. The destination workload is already at the active request limit. The moved request causes the workload to exceed the limit.
- A DBA moves an active request to a new workload. TASM has to override throttle limits to allow this action.
- A DBA forces a request in the delay queue to run immediately. TASM has to override throttle limits to allow this action.
- TASM changes the state, and the request limit for the new state is less than the limit for the prior state. The number of active requests exceeds the limit until some requests complete.
- TASM releases a delayed request that is causing a deadlock, even if the workload is already at the request limit.
- The DBA changes the ruleset. TASM counts active requests against the new ruleset, even if that causes the number of requests to exceed the new throttle limit.
- The Prevent Mid-transaction Throttle Delays option is enabled, and requests in a transaction execute when the throttle limit is reached.
It is also possible for throttle limits not to be met, even though there are requests in the delay queue, ready to execute, that are managed by that throttle. This can happen when a request is subject to multiple rules that perform throttling (including system throttles, workload throttles, and utility throttles), and one or more of them is already at the throttle limit. Because a request must satisfy the throttle limits for all applicable rules, if all rules cannot be satisfied, then the request is delayed.