Oracle Access Manager (OAM) How To Modify The CAPACITY CONSTRAINT (AccessCapacity) Parameter (Doc ID 2897837.1)

• A capacity constraint causes the server to reject requests only when it has reached its capacity. Capacity includes all requests, queued or executing, from the constrained work set. Work is rejected either when an individual capacity threshold is exceeded or if the global capacity is exceeded.
• The parameter "AccessCapacity" maintains the capacity that can be loaded onto the WebGate and any requests post this limit will be blocked which will ensure there is no hang at WebGate end. There is no metric/method to measure which is the inflight value. This parameter and its optimal value depends on individual topology (tuning).
• In the versions prior to 12c, the Weblogic CAPACITY CONSTRAINT was hard coded to unlimited, 10, and 300.
• In the OAM 12c versions, the Bundle Patch versions prior to OAM BUNDLE PATCH 12.2.1.3.210915 and OAM Bundle Patch 12.2.1.4.210920, the AccessCapacity is set to 10.

• When the capacity constraint is unlimited, the OAM server holds on to the requests from the WebGate (till AAA timeout happens), or whatever is blocking it on the server mitigates.
• When the capacity constraint is in place, the WebGate error is immediate, and allows the WebGate to reach out to other servers if they are available right away. The consequences of keeping the capacity constraint unlimited are more insidious.
• The servers will not reject requests even if the server has been overloaded. As a consequence, can end up with a server crash. This has been seen many times when the there is no further explanation to the failure other than Out of Memory (OOM), or overload, hence the change.

When CAPACITY CONSTRAINT is set to low ...

• Performance Degradation for Login
• "The WebGate plug-in is unable to  contact any Access Servers." in the logs.