Hi everyone,
A recent ServiceNow–Jira thread got me thinking about a problem that comes up in almost every two-way integration I've worked on (full disclosure, I work at ZigiWave): update loops.
A typical example:
- An incident is reopened in ServiceNow.
- The integration moves the linked Jira work item back to "In Progress".
- A Jira automation reacts to that transition and updates a field.
- That update syncs back to ServiceNow, which changes the incident again, and so on.
In my experience, the integration itself is rarely the cause. A second actor, like an automation rule, business rule or webhook, makes a follow-up change under a different account. Here's what has worked for us so far:
- A dedicated integration user on both sides. Filter out updates where the last updater is that user (Jira changelog author, ServiceNow sys_updated_by).
- Sync only real changes. Compare mapped fields with the last synced value instead of reacting to every update event.
- Make automations ignore the integration user. Status-based rules shouldn't fire when the integration made the change.
I'm curious how others handle this:
- Where do you usually put the loop check: in the integration, in Automation conditions, or in ServiceNow business rules?
- Does Jira "Done" close the ServiceNow record in your setup, or only propose a resolution that the ServiceNow owner confirms?
- Have you seen loops caused by something less obvious, like SLA updates, attachments or a mail handler?
Would love to hear what has worked (or failed) for you.