@PYaoIn this case there was second server in K2 environment to which some of the escalations processing was assigned. Server was removed from environment while escaltion items were still assigned to it. Solution was either to perform Go To operation for each instance affected by server removal, but there were thousand of instance it is probably easier to re-introduce server back to environment so that escalations processed and then remove it again if necessary making sure that nothing is assigned to this specific server.

As for increasing threads it is not relevant here and in general it is never a good idea unless you 100% sure that you understand why you increasing them and there is avsolutely no way to design your process/es and workload so that they are working within default tread pool limits. And even if latter is the case scale out (adding extra K2 server) is a recommended approach as opposed to increasing threads on existing server. In 90% of the cases increasing the thread pool limit to a higher value is just a workaround which hides a real issue which should be resolved in a different way.