For how long events are kept in the replication queue for the XDC?


We have a hopefully an easy question to answer.

In the scenario when the XDC is stood up and a secondary cluster goes under the maintenance events would be created on the primary but not collected by the secondary cluster.

For how long the not picked-up events will stay in the queue before they are cleaned up?

Does the queue metadata content have any play in the queue cleanup, e.g. in the scenario of the multiple clusters polling the events there could be a different sequence for each cluster due to the maintenance, replication lag etc.?


Tasks that havent been picked up should always stay in queue and i think this should not be issue for open workfows as their history is available. This behavior can be different for closed and deleted workflows wherre history is no longer available and im still checking on that behavior and will get back to you.