Agent to integrate Jest with ReportPortal.
- More about Jest
- More about ReportPortal
npm install --save-dev @reportportal/agent-js-jest
1. Create jest.config.js
file with reportportal configuration:
module.exports = {
testRunner: 'jest-circus/runner',
testRegex: ['/__tests__/.*.spec.js?$'],
reporters: [
'default',
[
'@reportportal/agent-js-jest',
{
apiKey: 'reportportalApiKey',
endpoint: 'https://your.reportportal.server/api/v1',
project: 'Your reportportal project name',
launch: 'Your launch name',
attributes: [
{
key: 'key',
value: 'value',
},
{
value: 'value',
},
],
description: 'Your launch description',
}
]
],
...
};
In case you use the jest config section of package.json
, add the following entry:
{
"jest": {
...
"reporters": [
"default",
["@reportportal/agent-js-jest",
{
"token": "reportportalApiKey",
"endpoint": "https://your.reportportal.server/api/v1",
"project": "Your reportportal project name",
"launch": "Your launch name",
"attributes": [
{
"key": "key",
"value": "value"
},
{
"value": "value"
}
],
"description": "Your launch description"
}]
],
...
}
}
The full list of available options presented below.
Option | Necessity | Default | Description |
---|---|---|---|
apiKey | Required | User's reportportal token from which you want to send requests. It can be found on the profile page of this user. | |
endpoint | Required | URL of your server. For example 'https://server:8080/api/v1'. Use api/v2 for asynchronous reporting. |
|
launch | Required | Name of launch at creation. | |
project | Required | The name of the project in which the launches will be created. | |
attributes | Optional | [] | Launch attributes. |
description | Optional | '' | Launch description. |
rerun | Optional | false | Enable rerun |
rerunOf | Optional | Not set | UUID of launch you want to rerun. If not specified, reportportal will update the latest launch with the same name. Works only if rerun set to true . |
mode | Optional | 'DEFAULT' | Results will be submitted to Launches page 'DEBUG' - Results will be submitted to Debug page. |
skippedIssue | Optional | true | reportportal provides feature to mark skipped tests as not 'To Investigate'. Option could be equal boolean values: true - skipped tests considered as issues and will be marked as 'To Investigate' on reportportal. false - skipped tests will not be marked as 'To Investigate' on application. |
debug | Optional | false | This flag allows seeing the logs of the client-javascript. Useful for debugging. |
launchId | Optional | Not set | The ID of an already existing launch. The launch must be in 'IN_PROGRESS' status while the tests are running. Please note that if this ID is provided, the launch will not be finished at the end of the run and must be finished separately. |
restClientConfig | Optional | Not set | The object with agent property for configure http(s) client, may contain other client options eg. timeout . Visit client-javascript for more details. |
isLaunchMergeRequired | Optional | false | This flag determines whether to create temp files with the UUIDs of started launches and allow them to be merged using client-javascript 's mergeLaunches method. Temp file format: rplaunch-${launch_uuid}.tmp . |
launchUuidPrint | Optional | false | Whether to print the current launch UUID. |
launchUuidPrintOutput | Optional | 'STDOUT' | Launch UUID printing output. Possible values: 'STDOUT', 'STDERR'. Works only if launchUuidPrint set to true . |
extendTestDescriptionWithLastError | Optional | true | If set to true the latest error log will be attached to the test case description. |
token | Deprecated | Not set | Use apiKey instead. |
The following options can be overridden using ENVIRONMENT variables:
Option | ENV variable | Note |
---|---|---|
apiKey | RP_API_KEY | |
project | RP_PROJECT_NAME | |
endpoint | RP_ENDPOINT | |
launch | RP_LAUNCH | |
attributes | RP_ATTRIBUTES | Format: key:value,key:value,value |
description | RP_DESCRIPTION | |
launchId | RP_LAUNCH_ID | |
mode | RP_MODE | |
token | RP_TOKEN | deprecated Use RP_API_KEY instead. |
This is for your convenience if you have a continuous job that runs your tests and may report results that point to a different reportportal project definition, launch name, or attributes.
2. Add script to package.json
file:
{
"scripts": {
"test": "jest --no-colors --detectOpenHandles --config ./jest.config.js"
}
}
The agent has support of retries. Read more about retries in jest.
This reporter provides ReportingApi
in global variables to use it directly in tests to send some additional data to the report.
Note: Run with the default test files concurrency may lead to inconsistent files attaching. ReportingApi
also does not support tests running in concurrent
mode at the moment.
We are going to fix this behavior in the future.
Send file to ReportPortal for the current test. Should be called inside of corresponding test.
ReportingApi.attachment(file: {name: string; type: string; content: string | Buffer;}, description?: string);
required: file
optional: description
Example:
test('should be passed with attachment', () => {
const fileName = 'test.png';
const fileContent = fs.readFileSync(path.resolve(__dirname, './attachments', fileName));
ReportingApi.attachment({
name: fileName,
type: 'image/png',
content: fileContent.toString('base64'),
}, 'Description');
expect(true).toBe(true);
});
The Jest supports test sharding on multiple machines.
Thus, in order to have a single launch in ReportPortal for sharded tests, additional customization is required. There are several options to achieve this:
Note: The @reportportal/client-javascript
SDK used here as a reference, but of course the same actions can be performed by sending requests to the ReportPortal API directly.
The complete example of launchId
usage with shards can be found for our examples repo with GitHub Actions pipeline, so you can use it as a reference while following this guide.
The agent supports the launchId
parameter to specify the ID of the already started launch.
This way, you can start the launch using @reportportal/client-javascript
before the test run and then specify its ID in the config or via environment variable.
- Trigger a launch before all tests.
The @reportportal/client-javascript
startLaunch
method can be used.
/*
* startLaunch.js
* */
const rpClient = require('@reportportal/client-javascript');
const rpConfig = {
// ...
};
async function startLaunch() {
const client = new rpClient(rpConfig);
// see https://github.com/reportportal/client-javascript?tab=readme-ov-file#startlaunch for the details
const response = await client.startLaunch({
name: rpConfig.launch,
attributes: rpConfig.attributes,
// etc.
}).promise;
return response.id;
}
const launchId = await startLaunch();
Received launchId
can be exported e.g. as an environment variable to your CI job.
- Specify the launch ID for each job. This step depends on your CI provider and the available ways to path some values to the Node.js process. The launch ID can be set directly to the reporter config.
/*
* jest.config.js
* */
const rpConfig = {
// ...
launchId: 'receivedLaunchId'
};
or just set as RP_LAUNCH_ID
environment variable.
With launch ID provided, the agent will attach all test results to that launch. So it won't be finished by the agent and should be finished separately.
- As a run post-step (when all tests finished), launch also needs to be finished separately.
The @reportportal/client-javascript
finishLaunch
method can be used.
/*
* finishLaunch.js
* */
const RPClient = require('@reportportal/client-javascript');
const rpConfig = {
// ...
};
const finishLaunch = async () => {
const client = new RPClient(rpConfig);
const launchTempId = client.startLaunch({ id: process.env.RP_LAUNCH_ID }).tempId;
// see https://github.com/reportportal/client-javascript?tab=readme-ov-file#finishlaunch for the details
await client.finishLaunch(launchTempId, {}).promise;
};
await finishLaunch();
This approach offers a way to merge several launches reported from different shards into one launch after the entire test execution completed and launches are finished.
- With this option the Auto-analysis, Pattern-analysis and Quality Gates will be triggered for each sharded launch individually.
- The launch numbering will be changed as each sharded launch will have its own number.
- The merged launch will be treated as a new launch with its own number.
This approach is equal to merging launches via ReportPortal UI.
- Specify a unique CI build ID as a launch attribute, which will be the same for different jobs in the same run (this could be a commit hash or something else). This step depends on your CI provider and the available ways to path some values to the Node.js process.
/*
* playwright.config.js
* */
const rpConfig = {
// ...
attributes: [
{
key: 'CI_BUILD_ID',
// e.g.
value: process.env.GITHUB_COMMIT_SHA,
}
],
};
- Collect the launch IDs and call the merge operation.
The ReportPortal API can be used to filter the required launches by the provided attribute to collect their IDs.
/*
* mergeRpLaunches.js
* */
const rpClient = require('@reportportal/client-javascript');
const rpConfig = {
// ...
};
const client = new rpClient(rpConfig);
async function mergeLaunches() {
const ciBuildId = process.env.CI_BUILD_ID;
if (!ciBuildId) {
console.error('To merge multiple launches, CI_BUILD_ID must not be empty');
return;
}
try {
// 1. Send request to get all launches with the same CI_BUILD_ID attribute value
const params = new URLSearchParams({
'filter.has.attributeValue': ciBuildId,
});
const launchSearchUrl = `launch?${params.toString()}`;
const response = await client.restClient.retrieveSyncAPI(launchSearchUrl);
// 2. Filter them to find launches that are in progress
const launchesInProgress = response.content.filter((launch) => launch.status === 'IN_PROGRESS');
// 3. If exists, just return. The steps can be repeated in some interval if needed
if (launchesInProgress.length) {
return;
}
// 4. If not, merge all found launches with the same CI_BUILD_ID attribute value
const launchIds = response.content.map((launch) => launch.id);
const request = client.getMergeLaunchesRequest(launchIds);
request.description = rpConfig.description;
request.extendSuitesDescription = false;
const mergeURL = 'launch/merge';
await client.restClient.create(mergeURL, request);
} catch (err) {
console.error('Fail to merge launches', err);
}
}
mergeLaunches();
Using a merge operation for huge launches can increase the load on ReportPortal's API. See the details and other parameters available for merge operation in ReportPortal API docs.
Note: Since the options described require additional effort, the ReportPortal team intends to create a CLI for them to make them easier to use, but with no ETA. Progress can be tracked in this issue.
Licensed under the Apache License v2.0
This code was based on the jest-junit and adapted by team members of Ontoforce for the ReportPortal upload. Ontoforce contributed this effort as Open Source to the ReportPortal project team.