How Automation Doctor Works
An overview of the architecture, what the app reads, what it never touches, and current known limitations.
What the app solves
Starting December 3, 2026, Atlassian begins billing for Automation Steps that exceed the monthly allowance of your plan — including Enterprise, which was previously unlimited. Jira does not natively show which specific automation rule is responsible for that consumption, or let you simulate a rule's cost before publishing it. Automation Doctor covers both gaps, inside Jira.
Architecture overview
Admin
| (1) Connects with an administrator email + API token
v
Automation Doctor (Forge app, runs inside Atlassian's infrastructure)
|
|--> api.atlassian.com/automation/... (rule definitions, public API)
|--> {site}.atlassian.net/gateway/.../highUsageFlows (observed usage, experimental)
|--> standard Jira REST API (JQL, changelog, fields) (structural estimate fallback)
v
In-product screens: Overview, Rules, Review Priority
The three screens
Overview
Sums observed Automation Steps consumption for every automation on the site, projects month-end as a range (not a single number), and compares it against the monthly allowance the administrator configures.
Rules
Lets you pick a specific automation and simulate its consumption, using observed data when available or a structural estimate (trigger, conditions, branches, loops) when it isn't.
Review Priority
A ranking that combines total consumption, execution volume, steps per execution, and allowance share, to point at which rule to review first.
What the app never does
- Never creates, edits, enables, disables, or deletes any automation rule.
- Never modifies issues, fields, comments, or any project data.
- Never sends notifications or interacts with end users outside the admin screen.
Data source reliability
| Source | Type | Reliability |
|---|---|---|
| Rule definitions (Automation Rule Management API) | Public, documented | High |
| Observed consumption (highUsageFlows / highestUsages) | Internal Atlassian UI endpoint | Experimental — not officially supported; the app degrades gracefully (Full / Execution-only / Unavailable) and falls back to the structural estimate |
| Structural estimate | Computed from the rule's own definition | High for structure; depends on the quality of the parameters entered |
Languages
Portuguese (Brazil), English (US), Spanish.
Support
Email support@app-rl.com.