Telemetry infrastructure

The repo provided scalable Telemetry infrastructure references for both Docker (viadocker-compose) and Kubernetes (via Helm) approaches. User could select any of them or combined both approaches that suite the real scenario for deployment, each sub-component/service shall be able to communicate via exposed ip and port to each other regardless of deployment approach.

The recommended setup of this repo as below:

Node::Telegraf Node-> Node::OpenTelemetry Collector Agent-> Backend::OpenTelemetry Collector Gateway-> Backend::InfluxDB-> Backend::Grafana

Deployment Order

Some sub-components/services have interdependencies, therefore, recommended to deploy them according to order below:

1. InfluxDB
   • Helm solution: ./helm/influxdb2
   • Docker solution: ./docker/influxdb
2. Open Telemetry Gateway (using gateway profile)
   • Helm solution: ./helm/otelcol
   • Docker solution: ./docker/otel
3. Open Telemetry Agent (using agent profile)
   • Helm solution: ./helm/otelcol
   • Docker solution: ./docker/otel
4. Grafana
   • Helm solution: ./helm/grafana
   • Docker solution: ./docker/grafana

Please read README in each respective directory for technical guidance.

Default port number exposed in each services

Service	Helm (via Kubernetes Node Port)	Docker-Compose
InfluxDB	http://{any node ip}:32701	http://{host ip}:58100
Grafana	http://{any node ip}:32601	http://{host ip}:59500
OpenTelemetry Gateway	http://{any node ip}:31082	http://{host ip}:52199
OpenTelemetry Agent	http://{any node ip}:30695	http://{host ip}:52199

How to connect Telegraf node to Telemetry Infrastructure setup

This is assumed that user already installed "telegraf" (installation).

typical telegraf run

telegraf --config demo.conf

enabling cpu and network metric for testing in Telegraf config file

....
[[inputs.cpu]]
    percpu = true
    totalcpu = true
    collect_cpu_time = false
    report_active = false
[[inputs.intel_powerstat]]
    cpu_metrics = ["cpu_frequency", "cpu_busy_frequency", "cpu_temperature", "cpu_c0_state_residency", "cpu_c1_state_residency", "cpu_c6_state_residency"]
...

connecting to Open Telemetry Gateway/Agent according to user setup in Telegraf config file

helm

[[outputs.opentelemetry]]
  service_address  = "{ip address of OpenTelemetry Gateway or Agent}:31082 or 30695"

docker-compose

[[outputs.opentelemetry]]
  service_address  = "{ip address of OpenTelemetry Gateway or Agent}:52199"

accessing Grafana Dashboard for data visualization

Open browser, paste following URL:

helm

http://{ip address of Grafana's host}:32601/

docker-compose

http://{ip address of Grafana's host}:59500/

Ensure browser's host is on reachable network to Telemetry Infrastructure host IP address.