Getting started

This repository is a sample application for users following the getting started guide at https://docs.docker.com/get-started/.

The application is based on the application from the getting started tutorial at https://github.com/docker/getting-started

Monitoring

An optional monitoring stack (Prometheus, Grafana, Loki, Promtail, Jaeger) is available in compose.monitoring.yaml to demonstrate metrics, logs, and traces for this app. It adds 5 extra containers (7 total), so it's heavier than the base tutorial. It's also deployed to the Azure VM, fronted by an nginx reverse proxy (compose.proxy.yaml) so the tools are reachable over the already-open port 80 without opening extra ports in the NSG.

The app exposes Prometheus metrics at /metrics and sends traces to Jaeger via OpenTelemetry auto-instrumentation (covering Express routes and MySQL queries). Logs stay on stdout and are scraped by Promtail directly from the Docker containers.

Launching

The monitoring and proxy stacks attach to the app-net network created by the main app stack, so start the app first:

# 1. Start the app stack — creates the shared "app-net" network
docker compose up -d

# 2. Start the monitoring stack — attaches to the existing app-net network
docker compose -f compose.monitoring.yaml up -d

# 3. Start the nginx reverse proxy — routes / to the app and /grafana/, /prometheus/, /jaeger/ to the tools
docker compose -f compose.proxy.yaml up -d

To stop, tear down in the reverse order (Docker refuses to remove a network that's still in use):

docker compose -f compose.proxy.yaml down
docker compose -f compose.monitoring.yaml down
docker compose down

Troubleshooting: if you start compose.monitoring.yaml or compose.proxy.yaml before compose.yaml, you'll see an error like network app-net declared as external, but could not be found. This is expected — it means app-net hasn't been created yet. Start the app stack first.

On the Azure VM, deploy-azure.yml runs all three docker compose stacks automatically in this order on every deploy.

URLs

Locally, each tool is still reachable directly on its own port:

Tool	URL	Notes
App	http://localhost:3000	direct container port (nginx also proxies it on port 80, see below)
Grafana	http://localhost:3001	login admin/admin; Prometheus/Loki/Jaeger datasources and a sample dashboard are pre-provisioned
Prometheus	http://localhost:9090	scrape targets and raw metrics
Jaeger UI	http://localhost:16686	search traces for service todo-app
Loki	http://localhost:3100	headless API; browse logs via Grafana Explore

Through nginx (port 80, either locally or on the VM at http://<vm-ip>/...), everything is reachable path-based on a single port:

Tool	URL
App	http://<host>/
App metrics	http://<host>/metrics
Grafana	http://<host>/grafana/
Prometheus	http://<host>/prometheus/
Jaeger UI	http://<host>/jaeger/

No authentication is enforced on Prometheus/Grafana/Jaeger — treat this as a training/demo setup, not a hardened production one.