diff --git a/METRICS.md b/METRICS.md
new file mode 100644
index 0000000..dde52a2
--- /dev/null
+++ b/METRICS.md
@@ -0,0 +1,107 @@
+# ZeaVis Edu — Metrics Endpoints
+
+This document lists every Prometheus metrics endpoint exposed by the ZeaVis Edu
+application stack and the payload each service provides.
+
+---
+
+## Overview
+
+| Service               | Host (prod)                       | Metrics Endpoint           | Port (local) |
+|-----------------------|-----------------------------------|----------------------------|--------------|
+| Web (Vite dev)        | `zeavisedu.asepharyana.my.id`     | `GET /metrics`             | 5173         |
+| API (Elysia)          | `api-zeavisedu.asepharyana.my.id`   | `GET /metrics`           | 3000         |
+| ML Service (Axum)     | `ml-zeavisedu.asepharyana.my.id`    | `GET /metrics`           | 8000         |
+| Prometheus Collector  | —                                 | `GET /metrics` (self)      | 9090         |
+
+> In production all metrics are scraped by the Prometheus collector running in the
+> Telemetry stack. See [`telemetry/prometheus/targets/`](./telemetry/prometheus/targets/)
+> for the auto‑discovery configuration.
+
+---
+
+## 1. Web App — `GET /metrics`
+
+| Endpoint          | Description                                      |
+|-------------------|--------------------------------------------------|
+| `/metrics`        | Vite dev‑server middleware + client‑side snapshot |
+
+### Metrics
+
+| Metric Name                         | Type    | Labels                        | Description                              |
+|-------------------------------------|---------|-------------------------------|------------------------------------------|
+| `zeavis_web_page_views_total`       | counter | —                             | Total page views this session            |
+| `zeavis_web_vital_bucket`           | gauge   | `name`, `rating`              | Last‑seen Web Vitals (CLS, FCP, INP…)   |
+
+**Development:** served inline by the Vite plugin `vite-plugin-metrics.ts`.
+**Production:** the static frontend serves no `/metrics` endpoint — consider
+forwarding the Vite dev server, or use the Telemetry collector to scrape
+client‑side beacons.
+
+---
+
+## 2. API (Elysia/Bun) — `GET /metrics`
+
+| Endpoint          | Description                                      |
+|-------------------|--------------------------------------------------|
+| `/metrics`        | Prometheus text format via `prom-client`         |
+
+### Metrics
+
+| Metric Name                                | Type      | Labels                         | Description                              |
+|--------------------------------------------|-----------|--------------------------------|------------------------------------------|
+| `zeavis_api_http_requests_total`           | counter   | `method`, `path`, `status`     | Total HTTP requests                      |
+| `zeavis_api_http_request_duration_seconds` | histogram | `method`, `path`               | Request latency buckets                  |
+| `zeavis_api_http_requests_active`          | gauge     | —                              | Concurrently‑handled requests            |
+| `zeavis_api_classifications_total`         | counter   | `result`                       | AI image classifications                 |
+| `zeavis_api_diagnoses_total`               | counter   | `disease`                      | Created diagnoses                        |
+| `zeavis_api_auth_operations_total`         | counter   | `operation`, `success`         | Login / register attempts                |
+| Default Node.js metrics                    | various   | —                              | CPU, memory, event‑loop lag, GC …       |
+
+**Source:** `apps/api/src/lib/telemetry.ts`, instrumented in `routes/`.
+
+---
+
+## 3. ML Service (Rust/Axum) — `GET /metrics`
+
+| Endpoint          | Description                                      |
+|-------------------|--------------------------------------------------|
+| `/metrics`        | Prometheus text format via `prometheus` crate    |
+
+### Metrics
+
+| Metric Name                                | Type      | Labels                         | Description                              |
+|--------------------------------------------|-----------|--------------------------------|------------------------------------------|
+| `zeavis_ml_http_requests_total`            | counter   | —                              | Total HTTP requests                      |
+| `zeavis_ml_http_request_duration_seconds`  | histogram | —                              | Request latency buckets                  |
+| `zeavis_ml_http_requests_active`           | gauge     | —                              | Concurrently‑handled requests            |
+| `zeavis_ml_predictions_total`              | counter   | —                              | Successful ONNX predictions              |
+| `zeavis_ml_model_load_status`              | gauge     | —                              | 1 = loaded, 0 = not loaded               |
+| Process metrics (libc/procfs)              | various   | —                              | RSS, CPU, fd count …                    |
+
+**Source:** `apps/ml-service/src/telemetry.rs`, instrumented in `routes.rs`.
+
+---
+
+## Prometheus Auto‑Discovery (Telemetry Stack)
+
+The Telemetry submodule includes a Prometheus instance that uses
+`file_sd_configs` to discover targets. Place a target file under
+`telemetry/prometheus/targets/` with content such as:
+
+```json
+[
+  {
+    "targets": ["zeavis-api:3000"],
+    "labels": { "service": "zeavis-api", "component": "backend" }
+  },
+  {
+    "targets": ["zeavis-ml:8000"],
+    "labels": { "service": "zeavis-ml", "component": "inference" }
+  }
+]
+```
+
+The Prometheus config (in `telemetry/prometheus/prometheus.yml`) will
+automatically pick up new files within its 15‑second scrape interval —
+no restart required.
diff --git a/docker-compose.yml b/docker-compose.yml
index 699f03f..7a70dcd 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -2,6 +2,9 @@ networks:
   app-shared-net:
     external: true
     name: app-shared-net
+  telemetry-net:
+    external: true
+    name: telemetry-net
 
 services:
   web:
@@ -10,6 +13,7 @@ services:
     restart: always
     networks:
       - app-shared-net
+      - telemetry-net
     env_file:
       - .env
     labels:
@@ -26,6 +30,7 @@ services:
     restart: always
     networks:
       - app-shared-net
+      - telemetry-net
     env_file:
       - .env
     environment:
@@ -47,6 +52,7 @@ services:
     restart: always
     networks:
       - app-shared-net
+      - telemetry-net
     env_file:
       - .env
     environment: