Merge pull request #5026 from krissetto/otel-test-stack

otel: simple OTEL collector/Prometheus stack for testing purposes
This commit is contained in:
Sebastiaan van Stijn 2024-05-02 10:28:47 +02:00 committed by GitHub
commit 647ccf3433
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 90 additions and 0 deletions

25
hack/otel/README.md Normal file
View File

@ -0,0 +1,25 @@
# Sample stack for testing OTEL functionality with the CLI
To test the OTEL functionality present in the CLI, you can spin up a small demo compose stack that includes:
- an OTEL collector container;
- a Prometheus container;
- an Aspire Dashboard container
The `hack/otel` directory contains the compose file with the services configured, along with 2 basic configuration files: one for the OTEL collector and one for Prometheus.
## How can I use it?
1) Start the compose stack by running `docker compose up -d` in the `hack/otel/` directory;
2) Export the env var used to override the OTLP endpoint:
`export DOCKER_CLI_OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317` (if running the CLI in a devcontainer or in other ways, you might have to change how you pass this env var);
3) Run the CLI to send some metrics to the endpoint;
4) Browse Prometheus at `http://localhost:9091/graph` or the Aspire Dashboard at `http://localhost:18888/metrics`;
5) In Prometheus, query `command_time_milliseconds_total` to see some metrics. In Aspire, select the resource in the dropdown.
> **Note**: The precise steps may vary based on how you're working on the codebase (buiding a binary and executing natively, running/debugging in a devcontainer, running the normal CLI as usual, etc... )
## Cleanup?
Run `docker compose down` in the `hack/otel/` directory.
You can also run `unset DOCKER_CLI_OTEL_EXPORTER_OTLP_ENDPOINT` to get rid of the OTLP override from your environment.

39
hack/otel/compose.yaml Normal file
View File

@ -0,0 +1,39 @@
name: cli-otel
services:
prometheus:
image: prom/prometheus:latest
command:
- "--config.file=/etc/prometheus/prom.yaml"
ports:
# Publish the Prometheus frontend on localhost:9091
- 9091:9090
restart: always
volumes:
# Store Prometheus data in a volume:
- prom_data:/prometheus
# Mount the prom.yml config file
- ./prom.yaml:/etc/prometheus/prom.yaml
aspire-dashboard:
image: mcr.microsoft.com/dotnet/nightly/aspire-dashboard:8.0-preview
ports:
- 18888:18888
environment:
DOTNET_DASHBOARD_UNSECURED_ALLOW_ANONYMOUS: 'true'
otelcol:
image: otel/opentelemetry-collector:latest
restart: always
depends_on:
- prometheus
- aspire-dashboard
ports:
- 4317:4317
volumes:
# Mount the otelcol.yml config file
- ./otelcol.yaml:/etc/otelcol/config.yaml
volumes:
prom_data:

20
hack/otel/otelcol.yaml Normal file
View File

@ -0,0 +1,20 @@
# Receive signals over gRPC and HTTP
receivers:
otlp:
protocols:
grpc:
http:
# Establish an endpoint for Prometheus to scrape from
exporters:
prometheus:
endpoint: "0.0.0.0:8889"
otlp/aspire:
endpoint: aspire-dashboard:18889
tls::insecure: true
service:
pipelines:
metrics:
receivers: [otlp]
exporters: [prometheus, otlp/aspire]

6
hack/otel/prom.yaml Normal file
View File

@ -0,0 +1,6 @@
# Configure Prometheus to scrape the OTel collector endpoint
scrape_configs:
- job_name: "otel-collector"
scrape_interval: 1s
static_configs:
- targets: ["otelcol:8889"]