Open olliesilvester opened 1 week ago
Tangential to this, as discussed with @DominicOram : Currently, when zocalo is staged, it waits for 2 seconds to wait for any stale messages to arrive so that they don't get used in the analysis. It would be cleaner if zocalo messages had an id unique to an instance of a plan.
With this, we wouldn't need to wait, and zocalo can just use the results with the correct id to process
We currently wait for the zocalo queue to clear after kicking off our gridscan. It's possible for zocalo to receive new results, and then clear them. Due to the speed of the CPU processing, we don't think we've seen this issue yet. However, as we are using much faster GPU processing soon, we should change this to make sure no data is lost
Acceptance Criteria