Overview
Traqory is an enterprise-grade, lightweight web analytics platform built for developers and digital teams who demand instant insights without the weight of invasive trackers. Designed from the ground up for minimal payload overhead and immediate aggregate queries.
The Problem
Most modern analytics tools are either bloated third-party scripts that slow down client websites or expensive enterprise platforms that sell user data. The challenge was building an ingestion engine capable of processing millions of un-sampled events without taxing client browsers or incurring huge cloud database costs.
The Solution
We engineered a dual-database architecture pairing ClickHouse for columnar event storage with PostgreSQL for account and metadata management. A lightweight (<1.2KB) tracking script streams events to an edge ingestion endpoint that batches writes directly to columnar tables.
Key Capabilities
- —Sub-1.2KB telemetry client with zero third-party dependencies
- —Real-time visitor sessions and conversion funnel tracking
- —Sub-50ms analytical query response time across millions of records
- —Cookie-free, GDPR and CCPA compliant by default
- —Custom goal tracking and UTM campaign attribution
Technical Decisions
Columnar Storage with ClickHouse
Relational databases struggle with ad-hoc analytical queries over millions of rows. ClickHouse provides 10x-20x data compression and processes columnar aggregations in milliseconds.
Edge Ingestion Buffer
Rather than opening individual database connections per incoming beacon, events hit an edge worker that buffers and flushes micro-batches every 500ms, preserving database health under traffic surges.
Server Components Dashboards
Rendered charts and data tables on the server using Next.js Server Components, drastically cutting client JavaScript bundle size while delivering instant first paints.
Architecture & Data Flow
Edge Telemetry Ingestion -> Memory Buffer Queue -> ClickHouse Event Warehouse (Aggregations) + PostgreSQL (Tenant & Config) -> Next.js Server-Rendered Analytics Dashboard.
My Contribution
Architected the full database schema, wrote the edge ingestion handler, built the dashboard UI with Next.js, and tuned ClickHouse table partitioning keys.
Outcome
Successfully scaled to handle tens of millions of events monthly with predictable response times, giving clients privacy-friendly visibility into their traffic.