Mirra Andreeva’s platform rests on a three‑tier architecture that separates user interaction, core processing, and data storage. This guide shows beginners how the layers fit together and how to start exploring each part.

Diagram showing the three‑tier architecture of Mirra Andreeva with presentation, application, and data layers.
Diagram showing the three‑tier architecture of Mirra Andreeva with presentation, application, and data layers.

What Is Mirra Andreeva?

Mirra Andreeva is a cloud‑native analytics service that lets users visualize real‑time metrics from multiple sources. It is designed for teams without deep engineering backgrounds, so the interface feels like a dashboard while the backend handles heavy data lifting.

The service is built on open‑source components, which keeps costs low and allows developers to replace parts without a full rewrite. Because of this modularity, new data connectors can be added in a few hours.

Flowchart of key components such as API gateway, service mesh, Kafka, Redis, and monitoring stack within Mirra Andreeva.
Flowchart of key components such as API gateway, service mesh, Kafka, Redis, and monitoring stack within Mirra Andreeva.

Layered Architecture Overview

The system follows a classic three‑tier model: Presentation, Application, and Data layers. Each tier communicates through well‑defined APIs, reducing the risk that a change in one layer breaks another.

In the Presentation layer, a single‑page web app built with React renders charts and tables. The Application layer hosts microservices written in Go and Python that perform aggregation, alerting, and user management. The Data layer stores raw events in a time‑series database and indexes summaries in a relational store.

LayerPrimary TechnologyKey Responsibility
PresentationReact + TypeScriptRender dashboards and handle user input
ApplicationGo microservices, Python workersProcess streams, apply business rules
DataInfluxDB, PostgreSQLStore time‑series events and relational metadata

Key System Components

The API gateway sits at the edge, routing external requests to the appropriate microservice and enforcing authentication. Service mesh tools like Envoy provide observability and traffic shaping between services.

A distributed cache (Redis) stores recent query results, cutting latency for repetitive dashboard loads. The data ingestion pipeline uses Kafka to buffer incoming events before they reach the processing services.

Finally, the monitoring stack—Prometheus for metrics and Grafana for visualization—keeps the platform itself observable, allowing operators to spot bottlenecks early.

Step‑by‑Step Exploration for Beginners

Follow these simple steps to get hands‑on with each architectural piece.

  1. Open the web UI and note the URL pattern for each dashboard. These URLs map to routes handled by the React router in the Presentation layer.
  2. Use the browser’s developer tools to inspect network calls. You will see calls to /api/v1/ endpoints, which are served by the API gateway.
  3. Clone the public GitHub repository (if available) and locate the services/ directory. Each folder contains a microservice written in Go or Python.
  4. Run docker-compose up to spin up the local stack, which includes Redis, Kafka, InfluxDB, and PostgreSQL. Observe how the services connect via environment variables.
  5. Open a terminal inside the ingestion/ container and publish a test event to Kafka. Then refresh a dashboard to see the new data appear, confirming the end‑to‑end flow.

These actions let you see how the Presentation, Application, and Data layers interact in real time.

Conclusion and First Action

Understanding Mirra Andreeva’s architecture helps you troubleshoot issues, extend functionality, and make informed scaling decisions. Today, clone the repository and run the local stack to witness the three‑tier flow yourself.

Frequently Asked Questions

What programming languages are used in Mirra Andreeva’s backend?

The backend primarily uses Go for high‑performance microservices and Python for data‑heavy workers.

Can I replace the time‑series database with another solution?

Yes, the architecture abstracts the database behind a storage interface, so swapping InfluxDB for Prometheus or TimescaleDB is supported with minimal code changes.

Is the platform suitable for on‑premise deployment?

While Mirra Andreeva is optimized for cloud environments, the Docker‑compose setup can be run on‑premise, provided you have sufficient resources for Kafka, Redis, and the databases.