Hi @diana
Thank you for your response.
Your reply reminded me that we had previously discussed request timeout issues a while back on this thread, where you warned us about the risks of being too “reserved” with our configuration. We needed to adjust the timeouts and increase the allowed body sizes on certain endpoints to allow the normal flow of traffic - especially due to the sync backlog.
Our investigation into this was supported by the metrics exposed by CHT. With this information, we were also able to handle a portion of these requests in memory, reducing I/O and the resulting delays.
For future reference, the feedback doc mentioned by Diana can be found in the user-specific -meta database, not the aggregated -meta or medic db. If I’m not mistaken, querying this space will require the DB admin credentials:
