Xray and Zephyr
With the Xray Cloud or Zephyr (Zephyr Scale, on Jira Cloud) integration, Proton automatically publishes run results to test management, creates the Tests for datasets, turns a Test Plan into an automation group and accepts runs started from Jira.
The screens use Xray's names (Test, Test Execution, Test Plan). When the organization uses only Zephyr, the texts switch to its names (Test Case, Test Cycle and Test Plan).
Set up
In Settings > Integrations, register the integration (see Integrations):
| Xray | Zephyr | |
|---|---|---|
| Base URL | The Jira Cloud where Xray is installed. | The Jira Cloud where Zephyr is installed. |
| Credentials | Client ID and Client Secret of an API key from Xray's global settings. | The Zephyr API Token (under "Zephyr API Access Tokens", in your Jira profile) and the Zephyr API region. |
| Projects | The Jira project keys (up to 50) and the default project. | Same. |
Publish a group's results
Publishing is set up per automation group. In the group, the Xray button in the header opens the Publishing to Xray drawer. In the Publishing tab:
| Field | What it is |
|---|---|
| Xray integration | Only shows up when there's more than one integration. |
| Project | The Jira project. |
| Test Plan | The plan the results go into. If empty, Proton creates a plan for the group. |
| Tests folder in Xray | Where the Tests created by Proton are stored. |
| Create the Test of datasets that have none | Proton creates the Test the first time the dataset is published. |
| Send the evidence when a run finishes | Sends the PDF report and the step screenshots. |
| Publish the group batches | Turns publishing on and off. |
After you save, the button shows the plan in use (or that publishing is paused). Stop publishing turns it off.
How publishing works
Publishing is automatic: each batch of runs in the group becomes a Test Execution, named "Proton · group · cycle · batch" and linked to the Test Plan. Each run sends its status, steps, linked bugs, a comment with the link to the run and, if requested, the evidence (up to 10 MB per file and 15 MB of screenshots per run).
- Proton groups runs that finish in sequence and publishes them within a few seconds.
- If Xray or Zephyr rejects the request, Proton tries again, with longer and longer intervals, up to 8 times.
- A dataset without a Test (with automatic creation turned off) is left out, and the screen tells you how many were.
The Batches tab shows each batch: Queued, the Test Execution key (which opens in Jira), Nothing to publish or the error with its reason. Publish again sends the batch one more time.
In the run details, the Test Execution, Test and Test Plan badges each open in Jira.
The Test of each dataset
Each dataset is linked to a Test. Besides the automatic creation during publishing:
- on the dataset card, No Test offers Create in Xray or Link to a Test that already exists;
- in the datasets list, select several and use Send to Xray to create their Tests all at once.
A Test created by Proton gets the steps from the components. In a Test that already existed, Proton doesn't change the steps. The issues the Test covers in Xray become requirements of the dataset in Proton. See Jira.
Import a Test Plan
In the Automation Center, Import from Xray creates an automation group from a Test Plan:
- Choose the plan: the project and the plan search.
- Review the Tests: enter the Group name and choose the Automation for the new datasets. For each Test in the plan, the screen shows what to do: Already linked, Link to a suggested dataset, Choose dataset, Create dataset or Don't import. The Linked, Suggested and No dataset filters help you review.
- Confirm. The group is created with a cycle and, with the Quality module, with the Test purpose.
The imported group shows which plan it came from, and Sync with the plan brings in the Tests added to the plan later. The import is done on desktop.
Run from Jira
The group can accept runs started in Jira, through a Jira Automation rule: for example, running the plan when a release moves to staging. The setup is in the drawer's Run from Jira tab, after publishing:
- Check Accept runs from Jira.
- Choose where it runs (runner or cluster), the Environment, whether to record video and whether to send an email when the batch finishes.
- Use Save.
The tab then shows The Jira Automation rule: the method, URL, headers and body of the request, each with a copy button. Build the rule in Jira with this data and with the API token of a robot account that can run in the Automation Center (the Executor profile can).
The rule calls Proton with the key of a Test Plan or a Test Execution, and Proton runs the datasets linked to its Tests. Last runs from Jira shows each request, Started or Refused with the reason (invalid key, no group accepts runs from Jira, cycle closed, no runner, repeated request in less than a minute, and others). A run from Jira counts as an immediate run.
In Zephyr, the Test Cycle and the Test Plan aren't Jira issues: the rule asks for the key or runs from an issue linked to a single Test Cycle.
Permissions and mobile
| Action | Profile permission |
|---|---|
| View and set up publishing and runs from Jira | Automation Group: read and update |
| Import and sync a Test Plan | Automation Group: create (and Dataset: create, to create datasets) |
| Create and link the dataset's Test | Dataset: read and update |
| The robot for runs from Jira | Automation Center: run |
The publishing drawer and the run details badges work on mobile; the Test Plan import doesn't.