Steps to reproduce:

1. Set up a service named 1008-5083 Bolafjall1 (dom-circuit) with the problem tag "Network contains 1008-5083". This is a child service of Mila-Network and Bola-Thve (dom-ring)
2. Status calculation rule: Most critical of child services
3. Service tag is circuit: 99
4. Trigger a problem that contains the tag "network: 1008-50832
5. Observe that the SLA is now 0 for the affected service and status is Major.
6. Clear the trigger, observe that the status is OK.
7. Observe the SLA counter - it should start increasing again...

Actual Result: counter stays at 0. How can the STATUS be OK but SLA be 0?

It appears that the SLA counter does not always restart when it should, following the recovery of a service.