Sambot's Observability Stack

CAUTION

This project is still nascent and might not work out-of-the-box in every scenario. This is experimental, use at your own descretion

What is this?

This ia a collection of tools and resources I am continusally experimenting with to help me understand how to build a robust observability stack for my projects. It is meant to be a living project that I optomize towards what I consider the best practices.

What is the stack?

--- #DRAFT from here ---

The stack is composed of the following components:

• Prometheus
• Grafana
• Loki
• Jaeger
• Thanos

Prometheus

Prometheus is a time series database that allows you to store and query time series data.

Grafana

Grafana is a dashboarding tool that allows you to visualize data from Prometheus.

Loki

Loki is a log aggregation system that allows you to store and query logs.

Jaeger

Jaeger is a distributed tracing system that allows you to trace requests across multiple services.

Thanos

Thanos is a Prometheus-compatible monitoring system that allows you to store and query metrics.

How do I use it?

I am currently using the following setup:

• Prometheus: https://prometheus.samast.dev
• Grafana: https://grafana.samast.dev
• Loki: https://loki.samast.dev
• Jaeger: https://jaeger.samast.dev
• Than

System Requirements

The core stack requires a single machine with the following (minimum) requirements:

• CPU: 4 cores
• RAM: 8 GB
• Disk: 20 GB
• Static IP

Deployment

The core stack requires a single machine at the moment so I use Docker Compose to deploy the stack.

To deploy the stack, run the following command:

docker-compose up -d

This will start all the services in the background.

To stop the stack, run the following command:

docker-compose down

This will stop all the services.

Data persistence

Backups

Hardening

---

References

• https://www.elastic.co/guide/en/elasticsearch/reference/8.17/docker.html#docker-compose-file