Sckoll
ActiveEmail infrastructure that scales.
Developer tool · Infrastructure · My role: Architected & Built
What it is
Sckoll is a high-throughput transactional email delivery infrastructure built in Go and TypeScript that provides multi-provider failover (Brevo, Cloudflare), queue-backed dispatch, and real-time deliverability telemetry.
Why it exists
Transactional emails (authentication OTPs, password resets, payment receipts) are mission-critical: if your primary delivery provider suffers an outage or flags an IP pool, user registration stops cold. Sckoll was built to ensure zero dropped transactional emails through intelligent multi-provider failover and local queuing.
What I did
- Engineered a high-performance Go dispatch engine utilizing buffered channels and worker pools for non-blocking local delivery
- Integrated multi-provider transport adapters (Brevo REST API, SMTP relays, and Cloudflare Workers) with automatic health probes
- Built token-bucket rate limiters and exponential backoff retry policies to prevent provider rate-limit rejections
- Developed an embeddable TypeScript client and REST API for drop-in replacement across microservices and Next.js applications
- Constructed real-time webhook ingestion for bounce handling, spam complaint tracking, and deliverability analytics
Technical decisions that mattered
Go concurrency over heavy message brokers
Instead of requiring a Kafka or RabbitMQ cluster for small-to-medium infrastructure footprints, Sckoll utilizes Go goroutines, buffered channels, and persistent SQLite write-ahead logging (WAL) for local queue reliability with single-binary deployment simplicity.
Circuit breaker with dynamic failover
When a primary provider returns 5xx errors or response latencies exceed 1500ms, the circuit breaker opens and immediately routes traffic to secondary and fallback transports without dropping requests or stalling calling services.
Strict DKIM/SPF/DMARC alignment validation
Pre-flight checks verify sender domain authentication before sending out emails, preventing domain reputation damage from misconfigured subdomains.
What I learned
Reliability in email infrastructure is fundamentally about handling degraded upstream states gracefully. Having an automated fallback path turns a critical production incident into an invisible background log entry.
Status
Active — infrastructure service