Overview
Validate a configuration, then run sample telemetry through it to check what it does.
Telflo checks a configuration in two ways.
| Check | Question it answers | Where |
|---|---|---|
| Validation | Is the configuration well-formed and valid? | Inline in the editor. See Validation. |
| Testing | Does the configuration process telemetry as expected? | The Testing tab, telflo test from the CLI, run_pipeline_test from the MCP server, or POST /testing/runs on the API. |
Validation is static: it checks structure and schema. Testing is dynamic: Telflo runs a real OpenTelemetry Collector of the configuration's distribution and selected version, feeds your sample telemetry through the configuration, and captures the output. A ClickStack configuration tested from the Testing tab runs on the OpenTelemetry Collector Contrib build its ClickStack version is based on. The same tests can be run from a terminal or CI with telflo test; see Pipeline tests from the CLI.
The Testing tab appears when testing is enabled for your account.
Testing tab
The Testing tab has three sub-tabs.
- Tests. Run the configuration against a sample input and review run history. See Running tests.
- Validation. Define rules that a run's output must satisfy. See Validating output.
- Suites. Group reusable tests and run them together. See Test suites.
The Testing tab requires a configuration version to be loaded; save a version first if prompted. A run tests the configuration as it currently appears in the editor.
What a test run does
- You select a sample input file and a pipeline.
- Telflo runs a collector that matches the configuration's collector distribution and version, for a set duration of 1 to 60 seconds. The duration starts at 5 seconds. For runs you start from the Tests sub-tab, the field afterwards remembers the value you last used in that browser; a run started from a suite uses the test configuration's saved duration, and a run started from the CLI or the API uses the duration in its spec or request. The distribution and version are fixed on the run when you start it.
- The sample telemetry is fed through the configuration, and the output is captured.
- The run is recorded in the history with a status, and you can inspect its output.
Tests outside the editor
A pipeline test can also be started from the command line with telflo test, or by any client through POST /testing/runs on the Telflo API. Both use the same sandbox as the Testing tab. The request names a stored configuration version, or sends YAML inline, together with a sample of up to 1 MB, the pipeline to test, a duration of 1 to 60 seconds, and either the rules to evaluate or an instruction to run the rules already saved on the version's Validation sub-tab.
A run that targets a stored version appears in that version's history on the Tests sub-tab, with cli-sample.json as its input file, and can be inspected there like any other run. A run on inline YAML belongs to no version and is read back through the CLI or the API only. The CLI and the API report a run with a failed rule as Fail, which the history shows as Validation Failed. $orgvar: references are not substituted in these runs. See Pipeline tests and API endpoints.
The following pages describe each part.
Last updated on