How to Self-Host Privacy-First Analytics with Plausible

The Privacy-First Web Analytics Imperative

As digital privacy regulations like GDPR and CCPA become more stringent, the tools used to monitor web traffic are undergoing scrutiny. Traditional platforms, such as Google Analytics, rely on tracking cookies and collect detailed user data, requiring complex cookie banners that degrade the user experience. Additionally, their heavy tracking scripts can slow down client-side rendering performance.

These factors have driven engineering teams to seek privacy-first, lightweight alternatives. Plausible Analytics is a popular open-source solution that provides web analytics without collecting personal data or using cookies.

Plausible is designed to compile aggregate visitor statistics, keeping script sizes small (under 1 KB) while maintaining compliance with privacy laws. In this guide, we will explore the steps to self-host Plausible Analytics on your own infrastructure using Docker Compose, configuring ClickHouse for analytical data storage, and setting up an Nginx reverse proxy.

To build a self-hosted analytics infrastructure, we must first understand Plausible's system architecture. Unlike simple database applications, Plausible uses a multi-component architecture to handle high volumes of traffic. The core web application is written in Elixir, leveraging the Phoenix framework to process incoming tracking requests with high concurrency.

Relational metadata, such as user accounts, site configurations, and subscription details, are stored in a PostgreSQL database. However, raw analytical event logs (like page views and custom goal triggers) are stored in ClickHouse.

ClickHouse is an open-source, column-oriented database management system designed for Online Analytical Processing (OLAP). By storing data in columns rather than rows, ClickHouse can query millions of event records in milliseconds, providing fast dashboard loads even as traffic scales.

Orchestrating Plausible with Docker Compose

To begin the deployment, we will construct a multi-container environment using Docker Compose. At Bramsley Digital Studio, we create a dedicated directory on our server and define a 'docker-compose.yml' file.

This file coordinates the interaction between the Plausible application, ClickHouse, PostgreSQL, and a Redis instance used for caching. Let's outline the essential structure of the compose configuration:

version: '3.3'
services:
  plausible_db:
    image: postgres:14-alpine
    volumes:
      - db-data:/var/lib/postgresql/data
    environment:
      - POSTGRES_PASSWORD=secure_postgres_pass

  plausible_events_db:
    image: clickhouse/clickhouse-server:22.3-alpine
    volumes:
      - event-data:/var/lib/clickhouse
      - ./clickhouse/clickhouse-config.xml:/etc/clickhouse-server/config.d/logging.xml

  plausible:
    image: plausible/analytics:latest
    command: sh -c "sleep 10 && /entrypoint.sh db createdb && /entrypoint.sh db migrate && /entrypoint.sh run"
    depends_on:
      - plausible_db
      - plausible_events_db
    ports:
      - "8000:8000"
    environment:
      - BASE_URL=https://analytics.yourdomain.com
      - SECRET_KEY_BASE=generate_a_random_long_string
      - DATABASE_URL=postgres://postgres:secure_postgres_pass@plausible_db:5432/plausible_db
      - CLICKHOUSE_DATABASE_URL=http://plausible_events_db:8123/plausible_events_db

This configuration defines the network links between the services. We include a delay in the Plausible startup command to ensure that PostgreSQL and ClickHouse are fully initialized before the application attempts to run database migrations, preventing startup crashes.

• Plausible Web Engine: Run on Elixir's Phoenix framework for high concurrent connection capacity.
• PostgreSQL Database: Stores configuration, user accounts, and site metadata.
• ClickHouse OLAP Database: Manages columns of raw analytical pageview and custom event data.

Optimizing ClickHouse for High-Throughput Analytics

Optimizing ClickHouse is a key step for ensuring system stability. Because ClickHouse is optimized for large batch writes rather than frequent small insertions, Plausible buffers incoming events in memory before flushing them to ClickHouse in batches. We can adjust ClickHouse's memory and log settings by providing custom configuration files.

For example, modifying the log rotation rules prevents ClickHouse from consuming excessive disk space with internal diagnostic logs. Additionally, we must monitor disk storage closely, as analytics data can grow rapidly on high-traffic sites. Configuring data compression and ensuring the underlying server uses SSD storage will improve query performance.

Configuring the Nginx Reverse Proxy

Once the containers are running, the next step is securing the installation using a reverse proxy and SSL certificates. We use Nginx to listen on port 80 and 443, proxying requests to the Plausible service running on port 8000.

Nginx handles SSL termination using Let's Encrypt certificates, which can be automated with Certbot. Here is a snippet of the Nginx configuration:

server {
    server_name analytics.yourdomain.com;

    location / {
        proxy_pass http://localhost:8000;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header Host $host;
    }
}

This setup forwards the client's real IP address and request headers to Plausible, which is necessary for accurate geographic detection and visitor calculations without storing identifiable data.

Tracking Script Integration and Dashboard Initialization

With the proxy active, we can log into the Plausible dashboard, register our website, and retrieve the JavaScript tracking snippet. The script is designed to load asynchronously, so it doesn't block DOM parsing. A common challenge in analytics is that tracking scripts are frequently blocked by ad-blockers, leading to underreported traffic.

To address this, Plausible allows you to proxy the tracking script through your own domain using Nginx rules. By making the tracking request appear as a local path (e.g., '/js/script.js' mapping to the Plausible script URL), you can bypass common blocking lists and collect more accurate visitor statistics.

Scheduling Database Backups and Updates

Maintaining a self-hosted analytics platform also requires planning for backups and updates. We configure cron jobs to run periodic database dumps of the PostgreSQL metadata and schedule backups of the ClickHouse directory.

When upgrading Plausible, we pull the latest Docker image, check the release notes for database schema changes, and restart the containers. This ensures the application remains secure and benefits from new features like custom event goals, API integrations, and email report delivery.

Analytics Pipeline Optimization at the Edge with Bramsley

Deploying a self-hosted, privacy-compliant analytics infrastructure is an effective way to optimize site performance and protect user privacy. However, managing ClickHouse clusters, configuring reverse proxies, and maintaining high availability can introduce operational overhead. At Bramsley, we architect and maintain custom analytics deployments at the edge.

We help enterprise brands configure privacy-first tracking pipelines, optimize ClickHouse databases, and set up edge proxy routes to bypass ad-blockers while maintaining compliance. Partner with us to implement analytics systems designed for speed and privacy.