Initial record retrieval slowness
Closed this issue · 4 comments
Description
To work out what Records the downloader needs to download, we currently pull a list of ALL records on an account. This request can be very slow especially on weekends. This is because we essentially ask Neos' cloud to give us a complete list of the records on your account which is often huge.
You can see this when you see the account downloader "stuck" on "Queueing Record Downloads"
Versions Affected
ALL, including ALL forks
Possible Solutions
Paginate the record search- Pagination already occurs from CloudX's RecordSearch class- Finish #21 (see notes on suitability in #21's discussion)
- Better status messages around the existing pagination.
Current Workarounds
Try any of the following
- Waiting - the request will eventually succeed
- Retrying Later - the request will be faster on non-peak times
Based on a report from @epicEaston197
RecordSearch is a CloudX class, Re-decompiling it, there is pagination in place.
We will continue research.
We could re-architect this to be more apparent about what's going on, but this really is a display/observability related issue rather than an actual bug.
This was resolved in #190