Docker Monitoring Stack

This directory contains a minimal Docker Compose setup for running:

• mtproto-zig
• Prometheus
• Grafana

The stack is intended for a VPS or any other Linux host where you want:

• the proxy exposed on a public TCP port
• the metrics endpoint available only on localhost
• Prometheus scraping metrics from the proxy container
• Grafana available over SSH port forwarding

Files

• docker-compose.yml
• prometheus.yml
• mtproto-zig-grafana.json

1. Prepare the proxy config

Create zigconf.toml next to docker-compose.yml.

The important part for metrics is:

[metrics]
enabled = true
host = "0.0.0.0"
port = 9400

host = "0.0.0.0" is required because Prometheus connects to the proxy from another container over the Docker network.

2. Set the proxy image tag

Edit docker-compose.yml and replace:

mtproto-zig:<change_me>

with the image tag you actually want to run.

3. Start the stack

From hack/docker run:

docker compose up -d

This starts:

• mtg-zig on 10000:8443
• Prometheus on 127.0.0.1:9090
• Grafana on 127.0.0.1:3000

Make sure that you have proper port in server.port configured. 443 is the default, but in that example we use 8443 as internal port inside container and 10000 as exposed port on the host.

4. Check that metrics are available

On the VM:

curl -s http://127.0.0.1:9400/metrics | head

Prometheus target check:

curl -s http://127.0.0.1:9090/api/v1/targets | jq .

You should see mtg-zig:9400 in the target list with health: "up".

5. Open Grafana over SSH

From your local machine:

ssh -L 3000:127.0.0.1:3000 <vm_ssh_user>@<vm_ip>

Then open:

http://127.0.0.1:3000

Default login from docker-compose.yml:

• username: admin
• password: admin

6. Add the Prometheus datasource in Grafana

In Grafana:

1. Open Connections -> Data sources
2. Click Add data source
3. Choose Prometheus
4. Set the URL to:

http://prometheus:9090

5. Click Save & test

This works because Grafana and Prometheus run in the same Compose network.

7. Import the dashboard

In Grafana:

1. Open Dashboards -> New -> Import
2. Upload mtproto-zig-grafana.json
3. Select your Prometheus datasource
4. Click Import

The dashboard includes:

• active connections
• memory usage
• total throughput
• total transferred bytes
• traffic throughput graphs
• per-user active connections
• top users by throughput
• top users by transferred bytes

Notes

• The proxy metrics are exposed on 127.0.0.1:9400 on the host, not publicly.
• Per-user metrics are generated from users in the proxy config loaded by the current process.
• Counter-based totals reset on proxy restart, which is expected for Prometheus counters.