* sqlstore: Run tests as integration tests
* Truncate database instead of re-creating it on each test
* Fix test description
See https://github.com/grafana/grafana/pull/12129
* Fix lint issues
* Fix postgres dialect after review suggestion
* Rename and document functions after review suggestion
* Add periods
* Fix auto-increment value for mysql dialect
Co-authored-by: Emil Tullstedt <emil.tullstedt@grafana.com>
(cherry picked from commit 4937f0daab)
* Alerting API: Send back 404 not found error for update and delete endpoints
* Alerting API: send back alert notification id for delete with uid endpoint
* AlertNotifications: Translate notifications IDs to UIDs in alert Rule builder
* Avoid shadowing errors, raise validation error on non-existing notification id
* create a cache for notification Uids to minimize db overhead
* add cache usage test
* avoid caching empty notification Uids
* isolate db in alert notificationUid caching tests
If notification state is pending and last update of state was made
less than a minute ago. In the case of a grafana instance is shut down/crashes
between setting pending state and before sending the notification/marks as complete
this logic should allow the notification to be sent after some time instead of
being left in an inconsistent state where no notifications are being sent.