Every Jira admin meets an automation loop sooner or later: a rule updates a field, another rule reacts, an integration syncs the change back, and suddenly your audit log has hundreds of executions and your service limit is gone. These are the five guardrails I check on every site I audit.
- Know who started it. Add a User condition "Initiator is not …" for integration and service accounts. It costs one line and prevents most loops with ServiceNow, Salesforce or any two-way sync.
- Keep "Allow rule trigger" off. In Rule details, this option lets other rules' actions trigger this rule. It's off by default for a reason. Turn it on only for rules designed to be chained, and document why.
- Use narrow triggers. "Work item updated" fires on every change. "Field value changed" with a specific list of fields fires only when it matters.
- Only write when the value actually changes. Before an Edit action, add a condition comparing the current value with the new one. Writing the same value still creates an update event for anyone listening.
- Watch the audit log weekly. Sort rules by executions. A rule that suddenly runs 10x more than usual is your early warning, before you hit the monthly limit.
Bonus: give each integration its own account, so you can filter it everywhere (Automation, JQL updatedBy, changelog).
What guardrails would you add? Share them in the comments.