Adding an agent (like Datadog) while deploying with GitHub

Troubleshooting and help
1

Summary:

  • In the interests of SEIM www.datadoghq.com cloud-siem, Security, and the Soc 2 Type 1 certification roadmap
  • We are considering deploying of Datadog agent docs.datadoghq /tracing/trace_collection/automatic_instrumentation/single-step-apm/?tab=linuxhostorvm to process and store logs, and provide additional application performance monitoring datadoghq product/apm/ (APM)

Ask:

  • Has anyone deployed the Datadog agent successfully on a Koyeb service?

  • Can we add an agent installation command to the build step? When I added the agent install from datadog to the ProcFile, it failed to sudo and apt update web: DD_API_KEY=$DD_API_KEY DD_SITE=$DD_SITE DD_ENV=$DD_ENV DD_APM_INSTRUMENTATION_ENABLED=host DD_APM_INSTRUMENTATION_LIBRARIES=python:3,js:5 bash -c "$(curl -L https://install.datadoghq.com/scripts/install_script_agent7.sh)" && streamlit run app.py --server.port $PORT --server.address 0.0.0.0

  • Any opinions welcome: Are we over-engineering the use case?

  • We are mindful of security and being proactive in due diligence. Koyeb provides the log exporter /docs/run-and-scale/log-exporter#custom-export-locations which we could configure, and it uses Datadog Vector vector.dev/docs/about/concepts/#sinks

AI says we should

Use datadogs official container image or datadog’s container-friendly installation methods.
Alternately:

  1. Use DogStatsD client only: Instead of installing the full agent
  2. Use Application Performance Monitoring (APM) without agent: You can use Datadog’s APM features by just installing the tracer:
  3. Container-optimized agent: If you absolutely need the full agent functionality, Datadog provides a container-optimized version that you might be able to run alongside your application (though this would require coordination with Koyeb’s platform capabilities)
1 Like