Skip to content

Latest commit

 

History

History
69 lines (49 loc) · 3.48 KB

README.md

File metadata and controls

69 lines (49 loc) · 3.48 KB

Hello World in OpenTelemetry

In this experiment we're going set up an environment to run the OpenTelemetry Collector manually. We'll use a simple configuration that does nothing (no operation). We'll also turn off any internal telemetry.

Tip: if you're using VS Code you can highlight the code snippets and use the command "Run Selection/Line in Active Terminal" to execute the commands.

Let's set up our test environment. We just need to install the tool "wget" to download the collector package. Please refer to the the Dockerfile for the full details.

We run the collector using the command /usr/bin/otelcol --config config.yaml. We'll discuss the config.yaml file later on

Let's build and run our test environment in Docker: ./start

We should see something similar in our terminal

2025-01-15T11:34:50.822Z        info    service@v0.117.0/service.go:164 Setting up own telemetry...
2025-01-15T11:34:50.823Z        info    service@v0.117.0/service.go:207 Skipped telemetry setup.
2025-01-15T11:34:50.830Z        info    service@v0.117.0/service.go:230 Starting otelcol-contrib...   {"Version": "0.117.0", "NumCPU": 2}
2025-01-15T11:34:50.830Z        info    extensions/extensions.go:39     Starting extensions...
2025-01-15T11:34:50.831Z        info    service@v0.117.0/service.go:253 Everything is ready. Begin running and processing data.

Our "Hello, World!" in this case is this log message "Everything is ready. Begin running and processing data.". Congratulations we've created a working OpenTelemetry Collector environment. Not a very useful one, but it's a good start. You can stop the collector by pressing Ctrl+C.

We can see the internal telemetry is disabled and the collector is ready to process data.

Let's review the config file to understand what's been configured.

For a working config, we need at least one receiver and one exporter. A useful component for both is the nop, this is short for "no operation". The nop receiver discards all data it receives and the nop exporter does nothing with the data it receives.

receivers:
  nop:

exporters:
  nop:

If you recall the output from the collector logs we could see that internal telemetry was skipped.

2025-01-16T08:08:02.979Z        info    service@v0.117.0/service.go:207 Skipped telemetry setup.

The default for the collector generates own telemetry (specifically metrics, logs and traces are handled differently) as Prometheus endpoint. To simplify this experiment we turned this off by setting the metrics level to none. We'll review this feature in a later experiment.

service:
  telemetry:
    metrics:
      level: none

Finally we need a way to connect the receiver and exporter together. This is done through a pipeline. As the name implies we can do some processing in between the receiver and exporter, something we'll look at later on.

  pipelines:
    logs:
      receivers: [nop]
      exporters: [nop]