blob: bb66e947c8e29abc725c1c0d1d329448c4de32c9 [file] [view]
# Plugin Test
Plugin tests are required and should pass before a new plugin is able to merge into the master branch.
Specify a support matrix in each plugin in the `skywalking/plugins` folder, along with their website links,
the matrix and links will be used for plugin support table documentation generation for this doc [Plugins.md](../setup/Plugins.md).
Use `make doc-gen` to generate a table and paste into `Plugins.md` after all test passes.
## SkyWalking Agent Test Tool (Mock Collector)
[SkyWalking Agent Test Tool](https://github.com/apache/skywalking-agent-test-tool)
respects the same protocol as the SkyWalking backend, and thus receives the report data from the agent side,
besides, it also exposes some HTTP endpoints for verification.
## Tested Service
A tested service is a service involving the plugin that is to be tested, and exposes some endpoints to trigger the instrumented
code and report log/trace/meter data to the mock collector.
## Docker Compose
`docker-compose` is used to orchestrate the mock collector and the tested service(s), the `docker-compose.yml` should be
able to run with `docker-compose -f docker-compose.yml up` in standalone mode, which can be used in debugging too.
## Expected Data
The `expected.data.yml` file contains the expected segment/log/meter data after we have triggered the instrumentation
and report to mock collector.
Once the mock collector receives data, we post the expected data to the mock collector and verify whether
they match.
This can be done through the `/dataValidate` of the mock collector, say `http://collector:12800/dataValidate`, for example.
## Example
If we want to test the plugin for the built-in library `http`, we will:
1. Build a tested service, which sets up an HTTP server by `http` library, and exposes an HTTP endpoint to be triggered in the test codes, say `/trigger`,
take this [provider service](https://github.com/apache/skywalking-python/blob/master/tests/plugin/http/sw_http/services/provider.py) as example.
2. Compose a `docker-compose.yml` file, orchestrating the service built in step 1 and the mock collector,
take this [docker-compose.yml](https://github.com/apache/skywalking-python/blob/master/tests/plugin/http/sw_http/docker-compose.yml) as an example.
3. Write test codes to trigger the endpoint in step 1, and send the expected data file to the mock collector to verify,
take this [test](https://github.com/apache/skywalking-python/blob/master/tests/plugin/http/sw_http/test_http.py) as example.