Status — EPFL SI

Note

This website is created from the EPFL's next.js starter kit

A simple next.js monitoring solution based on Prometheus.

Summary

• Status — EPFL SI
• Summary
• Requirements
• Getting Started
  • Clone the projet
  • Edit Alertmanager config
  • Launch Docker compose (dev kit)
  • Define environment variables
  • Launch Web Application (dev mode)
• Deploy with Kubernetes
  • Configure secrets
  • Deployment in test
  • Deployment in prod
• Usage of short-lived website

Requirements

• Bun ≥ 1.2
• A Microsoft Entra ID app registration with the following redirect URI: http://localhost:3000/api/auth/callback/microsoft-entra-id
• Access rights to the team epfl_status in keybase
• Be member of group "status-admins" in groups.epfl.ch

Getting Started

Clone the projet

First, to use the projet locally, you need to clone the projet Here the command below.

git clone git@github.com:epfl-si/status.git
cd status # Go to the root of the project

Edit Alertmanager config

First, you need to get secrets for email credentials.

Move to /devkit/alertmanager

Copy config.yml.example and rename it to config.yml

In this YML file, there is at the top, a "global" attributes with smtp variables as below

global:
  smtp_smarthost: ''
  smtp_auth_username: ''
  smtp_auth_password: ''
  smtp_require_tls: true
  smtp_from: '' # smtp_from have the same value as smtp_auth_username

Fill empty values with secrets from following files : /keybase/team/epfl_status/secrets.yml

Launch Docker compose (dev kit)

Now,to have Prometheus, Alertmanager and Blackbox Exporter containers, you need to move to the devkit folder and to start the docker compose as below.

cd devkit # Go to the /devkit
docker compose up -d # Start the docker compose

If you want to restart the Dev kit, just stop and restart the docker compose by using theses following commands

cd devkit # Go to the /devkit if you're not already
docker compose down # Stop the docker compose
docker compose up -d # Start the docker compose

Define environment variables

Before starting the web application, copy the .env.example and rename it .env.local.

cd .. # Go back to the root of the project
cp .env.example .env.local
vi .env.local # fill in .env.local using nano, vi, vim, or your favorite text editor

The file look like this

AUTH_SECRET=
ENTRA_ID=
ENTRA_SECRET=
ENTRA_ISSUER=https://login.microsoftonline.com/<tenant-id>/v2.0
PROMETHEUS_API_URL="http://localhost:9090"
ALERTMANAGER_API_URL="http://localhost:9093"
BLACKBOX_API_URL="http://localhost:9115"
ADMIN_GROUPS="status-admin_AppGrpU" # Separated by a comma without space, like "<group_name>_AppGrpU,<group_name_2>_AppGrpU"

Then, replace and fill variables with correct values.

Note

All theses secrets can be find to the team epfl_status folder in keybase (secrets.yml file)

Variable	Description
AUTH_SECRET	Change with a random secret for Auth.js session encryption (openssl rand -base64 32)
ENTRA_ID	Change with EntraID application (client) ID
ENTRA_SECRET	Change with EntraID client secret
ENTRA_ISSUER	Change <tenant-id> with your tenant id in `https://login.microsoftonline.com//v2.0**
PROMETHEUS_API_URL	Change only on test / prod environment. Keep it with devkit usage
ALERTMANAGER_API_URL	Change only on test / prod environment. Keep it with devkit usage
BLACKBOX_API_URL	Change only on test / prod environment. Keep it with devkit usage
ADMIN_GROUPS	Don't need to be changes since this group exist

Launch Web Application (dev mode)

Finally, you can install all dependencies and start the web application by using theses commands.

cd .. # Go back to the root of the project if you're not yet
bun install
bun dev

You can open and use the web application at http://localhost:3000.

Deploy with Kubernetes

Configure secrets

Duplicate twice secrets.yaml.exemple and name it secrets-test.yaml and secrets-test.yaml

replace all values betwwen "<" and ">" with test or prod values from keybase (/keybase/team/epfl_status/secrets.yml)

Deployment in test

First, you need to authenticate to OpenShift in your terminal

oc login --web --server=https://api.ocpitst0001.xaas.epfl.ch:6443

This command will open your navigator. You need to fill your credentials and after logged in, you can close your navigator. After that, if you're not in svc0041t-status-test project, you need to change it

oc project svc0041t-status-test

Next, push your secrets to OpenShift

kubectl apply -f secrets-test.yaml

To conclude, push the deployment configuration

kubectl apply -f deployment-test.yaml

Deployment in prod

First, you need to authenticate to OpenShift in your terminal

oc login --web --server=https://api.ocpitsp0001.xaas.epfl.ch:6443

This command will open your navigator. You need to fill your credentials and after logged in, you can close your navigator. After that, if you're not in svc0041p-status-prod project, you need to change it

oc project svc0041p-status-prod

Next, push your secrets

kubectl apply -f secrets-prod.yaml

To conclude, push the deployment configuration

kubectl apply -f deployment-prod.yaml

Usage of short-lived website

To test alerting, we're setup a "short lived website" that's able to be deploy and delete easily.

For this, there is a file called deployment-testserver.yaml.

First, you need to authenticate to OpenShift in your terminal

oc login --web --server=https://api.ocpitsp0001.xaas.epfl.ch:6443

This command will open your navigator. You need to fill your credentials and after logged in, you can close your navigator. After that, if you're not in svc0041p-status-prod project, you need to change it

oc project svc0041p-status-prod

To deploy the short-lived website, you need to execute the command below.

kubectl apply -f deployment-testserver.yaml

All you have to do for deleting is the command below.

kubectl delete -f deployment-testserver.yaml