Workflow approval task stuck in 'In Progress' and escalation not triggering

BPM AgilePoint
, , , , , , , ,
1

We have a critical issue with our approval workflow in AgilePoint NX Cloud. Multiple approval tasks are getting stuck in ‘In Progress’ status and never completing or escalating despite configured SLA rules.

The workflow is configured with:

  • 48-hour SLA for initial approval
  • Escalation to manager after 24 hours
  • Email notifications at escalation points

What we’re seeing:

  • Tasks remain ‘In Progress’ beyond the 48-hour window
  • No escalation emails are being sent
  • The escalation service appears to be running (checked service health dashboard)
  • Notification endpoints show as ‘Connected’ in Integration Hub

This is blocking critical purchase order approvals. We’ve verified the process model has correct SLA configuration and escalation paths defined. The notification service connectivity tests pass, but actual runtime escalations aren’t firing. Has anyone experienced similar escalation scheduling issues in the cloud environment?

2

Sarah, that’s really helpful. I checked and the service account permissions look correct, but I did find the SMTP issue - our cloud instance was trying to use port 25 which is blocked. However, even after fixing the notification connectivity, the escalations still aren’t triggering. The SLA timer seems to reset every time someone views the task. Is there a setting that prevents SLA progression when a task is being actively viewed?

3

We had the exact same problem after our cloud migration. The issue was twofold: First, the notification endpoints were configured but the cloud firewall rules weren’t allowing outbound SMTP on port 587. Second, the escalation service needs explicit permissions to modify task assignments in the cloud environment. Check your service account permissions in the AgilePoint Cloud Admin portal - the escalation service account needs ‘Process Instance Modify’ and ‘Task Reassignment’ permissions.

4

Mike, you’ve hit on a known behavior. By default, AgilePoint’s SLA engine can pause timers on ‘active engagement’ - when a user opens a task, the system assumes work is in progress. This is controlled by the ‘Pause SLA on Task Open’ setting in your process model’s advanced properties. However, this shouldn’t prevent escalation entirely, just delay it.

The real issue you’re facing is likely a combination of problems. Let me walk through the complete fix:

1. Escalation Service Scheduling: Verify in Process Administrator > System Settings > Escalation Service that:

  • Service is enabled and running
  • Polling interval is 5-10 minutes (not hours)
  • Service account has proper permissions (Process Instance Modify, Task Reassignment)

2. Process Model SLA Configuration: In your process model, open the approval activity properties:

  • Go to SLA tab and verify ‘Enable SLA’ is checked
  • Set SLA duration to 48 hours using ABSOLUTE time (not business hours unless you have a business calendar defined)
  • In Advanced SLA settings, set ‘SLA Time Base’ to ‘UTC’ (critical for cloud)
  • Disable ‘Pause SLA on Task Open’ or set a maximum pause duration of 2 hours
  • Configure escalation at 24 hours with action ‘Reassign to Manager’ and ‘Send Notification’

3. Notification Endpoint Connectivity: In Integration Hub > Connections:

  • Verify SMTP endpoint is using cloud-approved ports (587 for TLS or 465 for SSL)
  • Test the connection with ‘Send Test Email’
  • Check that notification templates reference the correct endpoint name
  • Verify cloud firewall allows outbound to your mail server IP

4. Migration-Specific Issues: Since you migrated from on-prem:

  • Re-publish your process model in the cloud environment (don’t just import the package)
  • Clear any cached SLA calculations: In Process Administrator, run ‘Reset SLA Timers’ for affected instances
  • Update any hard-coded server URLs in notification templates to use cloud endpoints

5. Debugging Steps:

  • Enable detailed logging for the escalation service (System Settings > Logging > EscalationService = Debug)
  • Monitor the logs during the next escalation window
  • Look for specific errors like ‘Notification endpoint unreachable’ or ‘SLA calculation error’

The combination of timezone misconfiguration, notification connectivity issues, and the ‘Pause on Task Open’ setting is likely causing your escalations to fail. After making these changes, create a test workflow with a 1-hour SLA to verify escalations work before applying to production processes.