Process Orchestrator

Escalations

Move overdue work to an escalation stage automatically, notify the owners, and manage the Escalation Monitor.

An escalation stage is the stage a process moves to when another stage, its parent, passes its due date without being completed. Use it for SLA breaches: a case not triaged in time goes to the support supervisor, an approval nobody answered goes to the approver’s manager. Escalation is built into the process; you do not need a scheduled flow to watch due dates.

Add an escalation stage

On the canvas, select Add escalation stage on the stage that can become overdue. Or open the stage, expand Escalation and select Add escalation stage.
A new stage named “New escalation stage for” followed by the parent’s name appears under the parent, joined by a dashed Overdue connector. It has an amber card with a clock badge.
Configure the escalation stage like any stage: Title, Stage Owner, Due Date, assignment and Completion Criteria.
Connect the escalation stage to the next element with at least one outgoing transition, for example back to the main path or to an End.

To remove it, open the parent’s Escalation section and select Remove escalation. The section shows Escalates to with the escalation stage, or No escalation. On an escalation stage, the section says which stage it belongs to: “This is the escalation stage of” its parent.

Rules

A stage can escalate to one escalation stage, and an escalation stage has exactly one parent.
An escalation stage has no incoming transitions; it is reached only by escalation.
The start stage cannot be an escalation stage, and a stage cannot escalate to itself.
Escalations can be chained (A escalates to B, B escalates to C) but cannot loop.
An escalation stage needs its own outgoing transition, like every stage.

What happens when a stage is overdue

A stage’s due date is the Due Date set in its Assignment section. When a stage with an escalation passes that date and is still active:

Its open tasks are canceled. Its On Exiting flows do not run.
The stage is marked Escalated.
The escalation stage starts like any stage: its On Entering flows run and its tasks are created. Its stage record’s Escalated From points back to the overdue stage visit.
Emails go to the owner of the overdue stage (“Overdue stage escalated: …”) and to the owner of the escalation stage (“New escalation stage: …”). If both stages have the same owner, one email is sent.

About the emails:

They are written in the recipient’s language (English or Arabic).
They are sent from the owner of the process instance (the Owner set when the instance started, not the owner of the process definition) when that owner is a user. For a team-owned instance, Dataverse’s default sender is used.
A team owner is reached through the team’s default queue.
A failure to send an email never stops the escalation.

A stage cannot be completed and escalated at the same time. Whichever happens first wins: if the last task is completed just before the escalation runs, the stage completes normally.

Stages without an escalation stage do not escalate. Their tasks simply stay open past their due date.

Escalation Monitor

Due dates are checked by the Escalation Monitor, a single record that the system creates and maintains. Find it at ProcessesEscalation Monitors.

Field	What it shows
Next Run On	When the monitor checks due dates next.
Last Run On	When it last ran.
Last Run Count	How many overdue stages its last run marked for escalation (at most 5,000 per run; any others are handled by the next run).
Is Idle	Yes when nothing is being watched.
Minimum Interval (minutes)	The shortest time between two runs. Defaults to 15; an empty, zero or negative value also means 15. You can change it.

How it schedules itself:

It plans its next run at the earliest due date among open stages that have an escalation, but never sooner than the minimum interval after its last run.
It goes idle when there is nothing left to watch, and wakes up when a stage with an escalation starts.
As a result, a stage can escalate up to one minimum interval after its due date. Lower the interval for tighter SLAs; raise it to reduce background load.

Users with the Process Designer role (and administrators) can view the monitor and change Minimum Interval (minutes). You do not create the record yourself.

Dynamics 365 CE on-premisesAsync Service requiredApplies to the other edition

The Escalation Monitor runs as Dataverse background processing. On Dynamics 365 CE on-premises, make sure the Asynchronous Processing Service is running.

Notifications without escalation

Apart from the escalation emails, Process Orchestrator sends no notifications of its own. To notify people when a stage starts or finishes, add a logic flow with a Send Email step to On Entering or On Exiting, or use Dataverse notifications on task assignment. For reminders before a due date, use a scheduled logic flow; see Schedules.

Example

The Triage case stage is due four business hours after a case is created. Its escalation stage Supervisor triage is owned by the support supervisors team and assigned with Team. If the agent has not completed triage in time, the agent’s task is canceled, the supervisors receive a team task and an email, and Supervisor triage continues to Investigate through a normal transition.

flowchart TD
  S([Start]) --> T[Triage case]
  T --> I[Investigate]
  T -. Overdue .-> U[Supervisor triage]
  U --> I
  I --> E((Resolved))