Claire_Conant
ServiceNow Employee
A task SLA can show numbers you didn't expect. You move an incident to On Hold and its SLA breaches anyway. The SLA reads Paused, but the breach flag is red. The breach time and original breach time show two different moments, and you have no idea which one belongs in a report. These all come down to one thing: a task SLA doesn't run a live clock. It calculates every number it shows from when the SLA started, the schedule it follows, and the time it spent paused. Once you know those three inputs, every surprising result has an explanation you can find.
 

When does the simple countdown model hold up?

 

It's natural to picture an SLA as a live countdown. For a simple 24/7 SLA with no pause conditions, that's about right—it starts when the record is created and runs straight to its deadline. But this changes as soon as a business schedule, a pause condition, or users in different time zones are involved, which in practice is most of the time.
 

How does the SLA engine calculate elapsed time?

 

A task SLA record is always in one of a few stages: In Progress, Paused, Completed, or Canceled. The stage tells you where the SLA stands, but it isn't what does the counting. When you update the task, the engine works out the elapsed time from three things: when the SLA started, which business schedule applies, and any time the SLA spent paused. It adds up the time that counts toward the SLA, and once that total reaches the duration on the SLA definition, the SLA breaches. So, the numbers on the record come from that calculation, not from a clock running in the background.
 

Why did the SLA breach while the incident was On Hold?

 

SLA definitions are the configured records that tell the platform how to run an SLA—its duration, its schedule, and its start, stop, and pause conditions. A pause condition is optional, but without it nothing pauses the SLA, no matter what the incident does.
 
For the SLA to pause, the pause condition has to recognize your On Hold state, for example, a condition that checks for State = On Hold. Miss that, and the SLA keeps counting straight through the hold and breaches. That's the number staring back at you on the record.
 

How is a business schedule different from a pause?

 

On the record, both can look alike, but they work in very different ways. A business schedule defines the hours an SLA counts—for example, Monday through Friday, 9:00 to 17:00. Any time that falls outside of those hours—overnight, weekends, holidays—is never counted toward the SLA at all. This isn't something that turns on and off; it's a standing rule built into the schedule.
 
A pause works differently. The SLA actively enters this state when a pause condition becomes true—for example, when an incident goes On Hold. A pause stops the clock during hours that would otherwise count, and the SLA's stage changes to Paused. The difference comes down to this: a schedule defines which hours ever count in the first place, while a pause stops the count during hours that normally would.
 

What's the difference between breach time and original breach time?

 

Breach time and original breach time look like the same value twice, until you see what each one is for. Both are estimates—the platform's calculation of when the SLA will breach—but they're made at different points and they update differently.
 
Original breach time is the first estimate. The engine calculates it once, when the SLA starts, by taking the start time, adding the SLA's duration, and mapping that across the schedule. After that, it doesn't change. Think of it as the breach date you were first quoted.
 
Breach time is the working estimate. It begins equal to original breach time, but it moves later each time the SLA pauses and resumes, because every pause pushes the finish back. Think of it as the revised date—like a delivery date that gets pushed back each time something delays the order. A gap between the two values is the running record of every pause the SLA has taken.
 
One more behavior applies after a breach. A system property—com.snc.sla.calculate_planned_end_time_after_breach—controls whether the platform keeps recalculating breach time once the SLA has already breached. When it's turned off, breach time is meant to stop moving at that point. So, if a post-breach value updates, or stays frozen, when you don't expect it, this property is usually why.
 

Why do users see different breach times?

 

Sometimes the times aren't really different at all. The breach fields store one underlying value, and each user sees it rendered in their own time zone. An admin in London and one in Sydney open the same SLA and see different breach times—both correct. Before you chase a bug that isn't there, check the time zone on each account.
 

How do you troubleshoot an SLA that looks wrong?

 

When an SLA looks wrong, check it in this order: the stage, then the pause condition on its definition, then the schedule that definition uses. For anything going into a stakeholder report, original breach time is the figure you can rely on, while breach time reflects the SLA's current estimate.
 
Several common SLA problems come straight from this same model:
  • An SLA that breaches a second after the record is created usually comes from the duration type on the SLA definition.
  • A breach that shows while seconds still remain usually occurs because the business elapsed percentage reaches 100 percent before the elapsed clock does.

 

Where to go from here

 

Once this model clicks, an SLA showing unexpected results becomes a puzzle you can solve. You read the stage, the schedule, and the pauses, and the result explains itself instead of leaving you to reverse-engineer a breach flag.

 

More info on this topic

 

Version history
Last update:
27m ago
Updated by:
Contributors