Activity execution limit restriction for process invocations

When you run a process that does not have an appropriate retry or loop breaker mechanism, the following error occurs:

You can review the process logs to find out the root cause of the activity limit breach.

You can resume the process execution after the 15-minute time limit expires or by republishing the process.

When you publish a process that goes into an infinite loop on a Secure Agent cluster, you might encounter a conflict in the behavior and not see the error message every time you run the process. This is because the in-memory cache used to store the process instances is node-specific and not broadcasted to all the nodes in the Secure Agent cluster. The system automatically chooses the nodes, and in a multi-node cluster setup, there might be cases where a process is rejected on one node but executed on another node. Each node maintains its own cache, allowing for independent handling of process failures and potential retries depending on the node where the process is executed.

Process A during its execution on node 1 encounters a 429- Too Many Requests error. As a result, an entry is added to the in-memory cache specific to node 1, indicating that process A failed due to an activity limit breach. The new instances of process A on node 1 will not be created for a period of 15 minutes.

However, if the process A execution is routed to node 2, node 2 will attempt to execute the process again. If process A encounters the same 429- Too Many Requests error on node 2, a new entry will be added to the cache specific to node 2.