What we do
Tools
Industries
Department
How we workLet's talk →
Home/Jira/Jira Automation Rules
Jira automation rules

Jira Automation Rules: Recipes You Can Copy, and the Limits That Stop Them

If you know your way around triggers and conditions, this page is the next step: rules for the jobs teams ask for most, the smart values that make them useful, the per-plan caps, and what the audit log is telling you when a rule stops. New to the rule builder? Start with our Jira automation guide, which builds a first rule from scratch.

Rule recipes for the jobs teams ask for

Each row is one rule. Atlassian's docs now call rules flows and issues work items, but the builder works the same.

The job Trigger, condition, action Smart values and catches
Share new bugs out across the team Work item created, condition type equals Bug, then Assign with the balanced workload or round-robin option over a list of people. Balanced workload counts open items per person, so someone back from leave gets the next few in a row.
Set a due date from priority Work item created, condition priority equals Highest, then Edit work item to set the due date. {{now.plusBusinessDays(2)}} skips weekends. Add one rule per priority, or an if/else block in a single rule.
Say who reopened something Work item transitioned from Done to In Progress, then Comment on it. {{initiator.displayName}} is the person who moved it. {{issue.assignee.displayName}} is who owns it now.
Start the parent when work starts Work item transitioned to In Progress, condition type is Sub-task, branch to the parent, condition parent status is To Do, then transition it. Without the status check, every later sub-task tries the same move and the audit log fills with errors.
A Monday digest of stale work Scheduled weekly, then Lookup work items with JQL such as updated <= -14d AND statusCategory != Done, then Send email to the lead. {{lookupIssues}} lists them, but only the first 100. Keep the JQL tight.
Ping chat on a Highest-priority bug Work item created or priority changed, condition priority equals Highest, then Send Slack message or Send Microsoft Teams message. {{issue.url}} gives a clickable link. {{issue.key}} alone isn't clickable in chat.
Let another system start a rule Incoming webhook trigger, then whatever actions the event calls for, such as create or transition a work item. Anything the caller sends arrives as {{webhookData}}. Keep the generated URL private: anyone holding it can start the rule.

What each Jira plan allows your rules

Steps are the monthly meter: each trigger, condition, action and branch that runs is one, pooled across your organization. Concurrency is how many rules can run at the same moment before the rest wait.

Plan Automation steps a month Rules running at once
Free 150 for the whole subscription 5
Standard 400 per user 10
Premium 750 per user 20
Enterprise 1,000 per user 30

The hard limits behind a rule that quietly stops

These are service limits, the same on every plan. Raising your step allowance doesn't move any of them.

The audit log says LOOP

How you'll know: Two rules edit the same field, or a rule reacts to its own change, and the run ends after about ten triggers in quick succession.

Fix: Add a condition that the field isn't already the value you're setting, or narrow which events the second rule listens to.

A scheduled rule misses items

How you'll know: The JQL in Jira's search returns 1,400 results, but the rule only ever acts on part of them.

Fix: A scheduled JQL search hands the rule at most 999 work items. Split it by project or date range, or process the full set through the API instead.

The rule switched itself off

How you'll know: It's disabled, and the last run touched a huge number of items.

Fix: One run can involve up to 5,000 associated items. Past that, Atlassian disables the rule. Break the job into smaller runs or move it off automation.

THROTTLED on a rule that calls another system

How you'll know: Runs that send web requests to a slow service start being held back.

Fix: Automation allows 60 minutes of processing time per 12 hours. Have the other system reply straight away and do its work afterwards.

Every rule on the site stopped at once

How you'll know: Nothing runs, even rules nobody touched, after a big bulk edit or import.

Fix: The global queue holds 50,000 items. A bulk change that fires thousands of rules can fill it. Pause noisy rules before big imports.

Fields are blank on a newly created item

How you'll know: A rule on work item created can't see fields that were filled in on create.

Fix: Add the Re-fetch work item data action at the start. It doesn't use a step.

You can't add one more component

How you'll know: The builder won't save a long rule.

Fix: A rule holds up to 65 components. Split it into two rules, or move the branching logic into code that calls Jira.

Reading a rule's audit log

  1. 1

    Open the log for that rule

    Space settings Automation the rule Audit log

    Global rules have the same tab under global automation. Each row is one run, newest first.

  2. 2

    Read the status

    SUCCESS did everything. NO ACTIONS PERFORMED stopped at a condition, which is often correct. SOME ERRORS ran but one action failed. FAILURE stopped on an error. LOOP and THROTTLED are the limits above.

  3. 3

    Expand the run

    You'll see each component, the work items it touched, and the error message from the one that failed, usually a permission, a missing field or a transition the workflow doesn't allow.

  4. 4

    Print what the rule saw

    Add a Log action with the smart values you're unsure of, such as {{issue.priority.name}} or {{webhookData}}, run it once, and read the value in the log.

    Watch for: the log action doesn't use any of your step allowance, so it's safe to leave in while you test.

Keep it as a rule, or move it into code?

Stay in automation

The job fits inside the limits

  • Fewer than 999 items per scheduled run, and well under 5,000 per run.
  • The actions stay in Jira, email or chat.
  • Someone on the team can edit it when the workflow changes.
Move it to the API

The job outgrew the builder

  • It sweeps thousands of items, or runs into LOOP and THROTTLED regularly.
  • It needs another system's answer before deciding what to do.
  • It eats a big share of your monthly steps for one repetitive task.

n-frames rebuilds rules that hit these limits as a small service listening to Jira webhooks and writing back through the Jira API, so the logic stays the same and the ceiling goes away.

Questions people ask

Is Jira automation free on every plan?

Yes, automation rules come with every Jira Cloud plan, including Free. What changes by plan is the monthly step allowance and how many rules run at once, as in the table above.

How many automation rules can I run in Jira?

You can create as many rules as you like. The limits are on what they do: steps per month for your plan, up to 30 running at the same moment on Enterprise, and 65 components inside any single rule.

Which triggers are available for Jira automation rules?

Work item created, transitioned, updated or deleted, field value changed, comment added, sprint and version events, scheduled, manual, and incoming webhook, among others. The recipes above use six of them.

Can a Jira automation rule send data to another system?

Yes, with the send web request action and any payload you build from smart values. You can tell the rule to wait for the reply and read it from {{webResponse}}. Atlassian doesn't retry a failed request later, so whatever receives it has to be reliable.

What are smart values in Jira automation?

Placeholders in double curly brackets that pull data into a rule, such as {{issue.summary}}, {{issue.assignee.displayName}} or {{now.plusBusinessDays(2)}}. They work in comments, emails, field edits, JQL and web request bodies.

Why does my Jira rule say LOOP or THROTTLED?

LOOP means rules kept triggering each other or themselves, about ten times in quick succession. THROTTLED means the rule used up its processing time, usually by waiting on a slow outside system. Neither is fixed by buying more steps.

Who rebuilds a Jira rule that has outgrown automation?

Usually whoever on the team can write and host a small service, which is the gap n-frames fills. We move the logic onto Jira's webhooks and REST API, run it for you, and keep the original rule's behavior, without the 999, 100 and 5,000 item caps.

Let's talk

Tell us the one thing your team does manually that eats up time. We read every message and reply within a day.