Provide the reason why an event was an expected UTD

[andybalaam]

Part of element-hq/element-meta#2638

Distinguish between 3 cases where we can't find keys for an old message:

• There is no backup -> HistoricalMessageAndBackupIsDisabled
• Backup is not working and we have not verified our device -> HistoricalMessageAndDeviceIsUnverified
• Something else -> Unknown

[codecov]

Codecov Report

Attention: Patch coverage is 92.85714% with 1 line in your changes missing coverage. Please review.

Project coverage is 85.28%. Comparing base (150d9e4) to head (dca0482).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/matrix-sdk/src/room/mod.rs	75.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4384   +/-   ##
=======================================
  Coverage   85.28%   85.28%           
=======================================
  Files         282      282           
  Lines       31196    31207   +11     
=======================================
+ Hits        26605    26615   +10     
- Misses       4591     4592    +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stefanceriu]

Super nice and clear, lgtm! 👏

[BillCarsonFr]

I think we need a 3rd variant to stay consistent with web, or we need to update web to be consistent with EX

[BillCarsonFr]

I think it's an improvement but I'd like to have some follow up tasks.

For example we fallback to Unknwown in determine_historical. It's a bit annoying because we report Unknwown in posthog as OlmKeysNotSent which sounds bad. This key was never sent. So maybe we need a BackupUnknown that would be mapped to something else in posthog (or we check the event age before reporting in posthog)

And just another comment, I though that we were caching the exists_on_server() but I don't see that in this PR?
Are we going to do a server query at ~every UTD report?

[andybalaam]

And just another comment, I though that we were caching the exists_on_server() but I don't see that in this PR? Are we going to do a server query at ~every UTD report?

Thank you for checking. Yes, that was done in #4356 so our call to exists_on_server in this PR uses the cached value where available.

Changes addressed, thanks!

[andybalaam]

I think it's an improvement but I'd like to have some follow up tasks.

For example we fallback to Unknwown in determine_historical. It's a bit annoying because we report Unknwown in posthog as OlmKeysNotSent which sounds bad. This key was never sent. So maybe we need a BackupUnknown that would be mapped to something else in posthog (or we check the event age before reporting in posthog)

Follow-up task is here: element-hq/element-meta#2666