send queue: mark ConcurrentRequestFailed as recoverable

[bnjbvr]

This will still disable the room's send queue, but the embedder may then decide whether to re-enable the queue or not, based on network connectivity. Ideally, we'd implement some retry mechanism here too, but since we're at a different layer than the HTTP one, we can't get that "for free", so let the embedder decide what to do here.

---

No test because the only way to trigger this is likely by having an encrypted room, start a /member query for that room, then try to send an encrypted message into that room. Sending encrypted messages require lots of endpoints, and I'm not sure whether we have all the testing setup for that; I'll ask on the Matrix room. But since this is a tiny change in behavior (and we're likely to have other tweaks like that related to error handling, I don't think each new error type marked as recoverable would need to require its own test).

---

Part of #3361.

[codecov]

Codecov Report

Attention: Patch coverage is 60.00000% with 2 lines in your changes missing coverage. Please review.

Project coverage is 84.19%. Comparing base (d41af39) to head (93a28f6).

Files	Patch %	Lines
crates/matrix-sdk/src/send_queue.rs	60.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3619      +/-   ##
==========================================
- Coverage   84.19%   84.19%   -0.01%     
==========================================
  Files         255      255              
  Lines       26416    26418       +2     
==========================================
+ Hits        22241    22242       +1     
- Misses       4175     4176       +1     

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

[poljar]

Makes sense, a bit annoying that we don't have access to all the erorr data we want.