We started to follow SLA values, day by day, and I found decrease in many IT services.

So, I started to try understand what were responsible were for this problem.

After some queries in the database, I noticed strange things in the SERVICEALARM table when checking servicealarmid and clock.fields (IT Service Detail SLA - Pool mdbotmdb018-otmdb2 instance otmdb2.png)

Look at that, sequentially servicealarmid 70849, happened on 2016-11-06 03:31-03-03, with value 4 (down) ....

E servicealarmid 70851 (greater than 70849), happened on 2016-11-06 03:30:59-03 (before 2016-11-06 03:31:00-03) with a value of 0 (UP).

I fixed this order just updating these two lines (Detail SLA IT Service - update command.png)

Now, the lines are following chronological ordem(Detalhe SLA IT Service - After update.jpg)

After these updates, SLA showed correctly.