LMS - Learning Management System

Quick Start

Clone the repo and run from the root of the project (where docker-compose.yml lives):

docker compose up --build

First run takes a few minutes to build the images. Subsequent starts are faster:

docker compose up

To stop everything:

docker compose down

To stop and wipe all data (database, Grafana dashboards, Prometheus data):

docker compose down -v

---

Local Development (without Docker)

Backend - also starts database:

cd LMS_API
dotnet run

API runs at http://localhost:8080

Frontend:

cd lms_frontend
npm install // if not installed already
npm run dev

Frontend runs at http://localhost:5173

---

Services

Service	URL	Description
Frontend	http://localhost	Vue.js app served by Nginx
API	http://localhost:8080	ASP.NET Core 9 REST API
API Metrics	http://localhost:8080/metrics	Prometheus scrape endpoint
Prometheus	http://localhost:9090	Metrics collection and storage
Grafana	http://localhost:3000	Dashboards and visualization
Node Exporter	http://localhost:9100	Host machine metrics
Loki	http://localhost:3100	Log aggregation and storage
Promtail	—	Log collector (no UI)

All the services and its ports can be found in docker-compose.yml.

Service Details

db — Microsoft SQL Server 2022. Seeded automatically on first start in Development mode. Data persisted in the sql_data Docker volume.

api — ASP.NET Core Web API. Depends on db. Will restart automatically while SQL Server is initialising (takes ~20s on first run). Exposes a /metrics endpoint for Prometheus.

web — Vue 3 frontend built with Vite and served via Nginx. Depends on api.

prometheus — Scrapes metrics every 15 seconds from the API and node-exporter. Config lives in prometheus.yml.

grafana — Visualisation dashboards. Default login is admin / admin. Data persisted in the grafana_data Docker volume. Prometheus and Loki are auto-provisioned as data sources via grafana/provisioning/datasources/datasources.yml.

node-exporter — Exposes host machine metrics (CPU, memory, disk, network) to Prometheus.

loki — Stores and indexes logs shipped by Promtail. Queried from Grafana using LogQL.

promtail — Tails Docker container logs and ships them to Loki. Config lives in promtail-config.yml.

---

Grafana Setup

Data sources (Prometheus and Loki) and dashboards are provisioned automatically on startup — no manual setup needed. After logging in (admin / admin), both data sources and the LMS dashboard folder will already be available.

Dashboards are loaded from grafana/provisioning/dashboards/. Any .json file placed there will be auto-imported on startup. To update a dashboard, export it as code from Grafana and overwrite the existing .json file, then restart the container.

Create a dashboard (if starting from scratch)

Dashboards → New → New dashboard → Add visualization

---

Grafana Dashboard Queries

API Metrics

Request rate (requests/second, by status code)

sum(rate(http_requests_received_total[5m])) by (code)

Error rate — 4xx and 5xx (%)

sum(rate(http_requests_received_total{code=~"4..|5.."}[5m])) / sum(rate(http_requests_received_total[5m])) * 100

Average response time (seconds)

rate(http_request_duration_seconds_sum[5m]) / rate(http_request_duration_seconds_count[5m])

Host Metrics (Node Exporter)

CPU usage (%)

100 - (avg(rate(node_cpu_seconds_total{mode="idle"}[5m])) * 100)

Memory used (%)

100 - ((node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100)

Log Queries (Loki / LogQL)

API logs (all)

{container="csharp_webapi"} |= ``

API logs (errors only)

{container="csharp_webapi"} |= `error`

Frontend logs (all)

{container="vue_frontend"} |= ``

Frontend logs (errors only)

{container="vue_frontend"} |= `error`

Grafana Alerts

Create them under A graph card → More → New alert rule. All alerts go in folder LMS, group LMS, evaluation interval 1m, keep firing none.

Avg Response Time

Query:

rate(http_request_duration_seconds_sum[5m]) / rate(http_request_duration_seconds_count[5m])

Condition: IS ABOVE 0.5 (500ms) Summary: High average response time on LMS API Description: Average response time has exceeded 500ms for more than 5 minutes.

---

HTTP Error Rate

Query:

sum(rate(http_requests_received_total{code=~"4..|5.."}[5m])) / sum(rate(http_requests_received_total[5m])) * 100

Condition: IS ABOVE 5 (5%) Summary: High HTTP error rate on LMS API Description: HTTP error rate (4xx/5xx) has exceeded 5% for more than 5 minutes.

---

CPU Usage

Query:

100 - (avg(rate(node_cpu_seconds_total{mode="idle"}[5m])) * 100)

Condition: IS ABOVE 80 (80%) Summary: High CPU usage on LMS host Description: CPU usage has exceeded 80% for more than 5 minutes.

---

Memory Usage

Query:

100 - ((node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100)

Condition: IS ABOVE 85 (85%) Summary: High memory usage on LMS host Description: Memory usage has exceeded 85% for more than 5 minutes.

---