ACRA
Bugs are keyed by a hash on the server
This simplifies code dealing with previously compound keys. It easily allows single bugs to be retrieved directly. It makes the 'resolved' documents smaller.
In the storage database view indexing is faster and smaller.
This is required to work with acra-storage PR ACRA/acra-storage#6
Btw, just FYI, you don't you don't need to recreate, you can just push (even forced) again, it'll update the code.
(Thanks for the tip on the forcing!)
More dashboard changes that work with the acra-storage changes I added to the PR there: https://github.com/ACRA/acra-storage/pull/6
I could see the stale data change being awkward for a very low traffic Acralyzer server. Should it be configurable? Should a stale request and a non-stale request with a timeout be launched at the same time? (A bit like an iOS app where you see old data from a local database after opening it back up, while the app is trying to contact the server to refresh.)