How to run a scan
How to run a scan
Running a scan takes less than a minute. This page covers what you need before you start, the steps for Scan Your Environment (Environment Analysis), and what the scan reads, returns, and keeps private.
Before you start
There is no API token to configure, and no credentials to paste. LaunchPad is a Forge app you install from the Atlassian Marketplace. When you install it, you consent to the permissions it needs once, and after that the scan runs using your own Jira identity. There is nothing else to set up.
You do need two things to run a scan:
- Jira administrator rights. Environment Analysis is restricted to admins. If you are not in an admin group, the scan returns an authorisation error rather than running.
- The app installed. Environment Analysis itself is free in every licence state, so you can run it whether or not you hold a paid licence.
The scan is read-only. It only sends GET requests, so it cannot change anything in your instance.
Running the scan
- Open LaunchPad from your Jira instance.
- Click Scan Your Environment on the home screen. (You can also use the Scan your environment hero button, or press S.)
- Click Run analysis.
That's it. The scan typically takes three to five seconds.
A terminal sidebar shows progress as LaunchPad reads your instance. When the scan completes, your results appear automatically.
How long it takes
The scan is fast, typically 3 to 5 seconds. There's nothing to wait for. Results appear immediately.
On larger instances the scan runs in several short passes behind the scenes (it reads in stages and saves its progress between each one) so it never runs into Jira's function time limit. The terminal sidebar shows which pass is running, and the results still appear automatically when it finishes.
Permissions
Two things govern who can run a scan and what it can read.
- You must be a Jira administrator. The scan checks your group membership before it does anything, and stops with an authorisation error if you are not an admin.
- The app's permissions cover the reads. Environment Analysis does not ask for any extra access beyond what LaunchPad already requested at install. It reads service desks, projects, custom fields, and request types using the app's existing Jira and Jira Service Management read permissions, all of which you granted when you installed the app.
If LaunchPad cannot read part of your instance (for example, a project it does not have access to), the scan does not fail. It carries on and shows a warning so you know that part of the picture may be incomplete.
What the scan covers
A single Environment Analysis scan looks across your whole Jira Service Management environment. It reads five sources of signal:
- Service desks in your instance.
- Projects across your instance.
- Custom fields defined in Jira.
- Request types on each service desk.
- The fields on each request type.
It does not deploy anything, and it does not read ticket contents.
What data is analysed
The scan only reads configuration metadata, the names and descriptions of the things above. It does not read the bodies of tickets, comments, or any end-user personal data.
From that metadata, LaunchPad works out what kind of service model your environment looks like it needs. It breaks the names and descriptions into words and word pairs, normalises spelling differences (including British and American spellings), filters out noise such as default sample projects, and matches what is left against the keyword families for each of its service models. Stronger and more varied signals score higher, and a few guard rules hold a recommendation back when the evidence is thin.
Privacy
LaunchPad is built to read as little as it can and to keep it inside Atlassian.
- No credentials are stored. There is no API token to enter, so there is nothing for LaunchPad to keep. Every call the scan makes uses Forge's built-in app and user authentication.
- No ticket content or personal data is read. The scan reads configuration metadata only (names and descriptions of desks, projects, fields, and request types), never the contents of requests or comments.
- Scan state stays in your Forge storage. While a scan runs, LaunchPad saves its progress and the final result in the app's own Forge storage inside your Atlassian site. Each new scan replaces the previous result.
What you get back
When the scan finishes, the results panel shows:
- Recommended service models. A ranked shortlist (up to five) of the service models that best fit what the scan found, each marked High, Medium, or Low confidence. Each recommendation has an Install this template action that takes you to deploy it.
- Suppressed recommendations. Any models that a stronger match outscored are tucked into a separate, collapsible section so the main list stays clear. You can expand it to see them.
- A signal summary. Counts of the service desks, request types, request-type fields, and custom fields the scan looked at.
- Integration suggestions. Where the signals point that way, suggestions for tools such as CSV Import or partner integrations that could feed your service model.
- Warnings. Dismissible banners (and a Warnings entry in the terminal sidebar) for anything the scan could not fully read, for example a project it lacked permission to see.
Environment Analysis recommends which service model to deploy. If you want to improve a service model you have already deployed (for example, spotting fields that could use an Assets object picker, or duplicate fields across request types), that is the job of the Request Flow Analysis scanner, which runs per project.
What to do next
Once you have results, the next step is understanding what they mean. See Understanding scan results for a guide to reading and interpreting what the scanner found.