Setup PostHog on AWS Spot Instances: Complete Guide

Required Tools and Dependencies

You'll need the AWS CLI configured with appropriate credentials, Docker and Docker Compose for container deployment, and basic knowledge of Linux system administration. Install these tools locally for deployment management. Additionally, having a domain name and DNS access is recommended for setting up SSL/TLS certificates with Let's Encrypt.

Recommended Instance Types for PostHog

PostHog's resource requirements depend on your event volume. For low-traffic deployments (under 10K events/day), a t3.large (2 vCPU, 8GB RAM) is sufficient. Medium traffic (10K-50K events/day) needs t3.xlarge or m5.large (4 vCPU, 16GB RAM). High-volume production deployments (over 200K events/day) benefit from m5.2xlarge (8 vCPU, 32GB RAM). Configure your Spot Fleet to include multiple instance families for maximum availability.

For the most up-to-date and detailed prerequisites, please refer to the official PostHog documentation.

Configuring Security Groups and Networking

Create a security group that allows inbound traffic on port 80 (HTTP), 443 (HTTPS), and 22 (SSH) from your IP address. PostHog runs on port 8000 by default, but you'll proxy it through nginx. Allow outbound traffic to all destinations so PostHog can fetch updates and connect to external services.

For production deployments, use a VPC with private subnets for the database layer and public subnets for the application layer. Configure NAT gateways to allow private instances to access the internet for updates. This network architecture provides better security while maintaining functionality.

Setting Up Persistent Storage with EBS

Create an EBS volume (minimum 100GB, gp3 type recommended) in the same availability zone as your Spot Instance. This volume will store PostgreSQL and ClickHouse data, ensuring your analytics data persists across Spot Instance interruptions. Mount this volume at `/mnt/posthog-data` and configure your Docker volumes to use subdirectories.

Configure your user data script to automatically attach and mount the EBS volume on instance launch. Use the volume's ID in your script and implement proper mounting with fstab entries. Enable automated snapshots through AWS Backup or Data Lifecycle Manager to create daily backups. This ensures you can recover your data even if both the instance and volume are lost.

Deploying PostHog Container

Download the official PostHog Docker Compose file from their GitHub repository. This file includes PostHog, PostgreSQL, Redis, and ClickHouse containers properly configured to work together. Customize the compose file to point volume mounts to your persistent EBS mount point.

Launch PostHog with `docker compose up -d` to start all containers in detached mode. Monitor the logs with `docker compose logs -f` to ensure all services start correctly. The first launch takes 2-3 minutes as PostHog initializes databases and runs migrations. Once complete, PostHog will be accessible on port 8000.

version: '3'
services:
  posthog:
    image: posthog/posthog:latest
    ports:
      - "8000:8000"
    environment:
      DATABASE_URL: postgres://posthog:posthog@db:5432/posthog
      REDIS_URL: redis://redis:6379
      CLICKHOUSE_HOST: clickhouse
    volumes:
      - /mnt/posthog-data/media:/code/media
    depends_on:
      - db
      - redis
      - clickhouse
  
  db:
    image: postgres:14
    volumes:
      - /mnt/posthog-data/postgres:/var/lib/postgresql/data
    environment:
      POSTGRES_USER: posthog
      POSTGRES_PASSWORD: posthog
      POSTGRES_DB: posthog
  
  redis:
    image: redis:7
    volumes:
      - /mnt/posthog-data/redis:/data
  
  clickhouse:
    image: clickhouse/clickhouse-server:23
    volumes:
      - /mnt/posthog-data/clickhouse:/var/lib/clickhouse

Configuring PostgreSQL and ClickHouse

PostgreSQL stores PostHog's application data (users, dashboards, feature flags) while ClickHouse handles analytics events for high-performance queries. Both databases are configured automatically by the Docker Compose file, but you should tune them for production use.

For PostgreSQL, increase shared_buffers and effective_cache_size based on your instance memory. For ClickHouse, configure max_memory_usage to use about 80% of available RAM minus what PostgreSQL needs. These optimizations ensure efficient query performance as your event volume grows. Monitor database disk usage and configure log rotation to prevent storage issues.

Data Persistence Strategy

Your PostHog data persistence strategy relies on three layers: EBS volumes for primary storage, automated EBS snapshots for point-in-time recovery, and optional S3 backups for long-term retention. Configure daily automated snapshots through AWS Backup or Data Lifecycle Manager policies.

Test your recovery procedure by launching a new Spot Instance and attaching a snapshot-restored volume. Verify PostHog starts correctly with all historical data intact. Document your recovery process and keep snapshots for at least 30 days to handle any data corruption scenarios.

Automatic Instance Replacement

Configure your Spot Fleet to automatically launch replacement instances when one is terminated. The new instance should automatically attach your persistent EBS volume and restart PostHog containers through user data scripts. This automation ensures minimal downtime—typically under 2 minutes.

Implement a startup script that checks for the PostHog data volume, mounts it if present, and launches Docker Compose automatically. Add health checks to verify PostHog is responding before marking the instance as healthy. Configure CloudWatch alarms to notify you of instance replacements so you can monitor interruption frequency.

Setting Up SSL/TLS with Let's Encrypt

Install certbot and obtain a free SSL certificate from Let's Encrypt for your PostHog domain. Configure nginx as a reverse proxy in front of PostHog to handle SSL termination. This provides encrypted connections and proper HTTPS URLs for your analytics dashboard.

Set up automatic certificate renewal with a cron job running `certbot renew`. Configure nginx to redirect HTTP to HTTPS and set appropriate security headers. Test your SSL configuration with SSL Labs to ensure it meets security best practices. Proper SSL setup is essential for collecting events from web applications.

Performance Tuning for Cost Efficiency

Optimize PostHog's performance to run efficiently on smaller (cheaper) instances. Configure data retention policies to automatically delete old events, reducing storage costs and query times. Set session replay retention to 30-60 days unless you need longer periods for compliance.

Enable ClickHouse's data compression and configure appropriate TTL policies for different event types. Monitor query performance and add indexes for commonly filtered properties. These optimizations can reduce your instance size requirements by 30-50%, multiplying your Spot Instance savings. Regularly review your performance metrics and right-size your infrastructure.

Backup and Recovery Procedures

Implement a comprehensive backup strategy with automated daily EBS snapshots, weekly full database exports to S3, and tested recovery procedures. Document your recovery time objective (RTO) and recovery point objective (RPO) to set appropriate backup frequencies.

Test your recovery procedure quarterly by restoring a backup to a separate environment and verifying data integrity. Keep at least 30 days of snapshots and 90 days of S3 backups. Configure lifecycle policies to automatically delete old backups and control storage costs. A well-tested backup strategy protects against data loss from any source, not just Spot Instance interruptions.