Lessons learned from running a SaaS entirely on Swift Vapor
summary
A Swift Vapor backend powers Telemetry Deck, originally a private sexual-health analytics app that grows to serve 14 million users per month; events are anonymized on-device, hit an ingestion API, queue through a message broker and land in Apache Druid for time-series queries, surfaced by a Swift API and a web UI. Operational stack uses Grafana, Loki and BetterStack for alerts; Redis/Valky caching cost lessons force TTLs; custom queuing is replaced by Vapor queues; Docker layer caching cuts builds from ~10 minutes to ~1.
AI-written summary · about
chapters (9)
- 0:00Introduction and project origin
- 5:56Nine lessons from Vapor experience
- 12:29API versioning and DTO management
- 19:18Caching strategies and pitfalls
- 22:31Queues and background jobs
- 24:59External services and SDKs
- 28:15Database migrations and optimizations
- 30:01Docker build time improvements
- 31:24Summary and project outcomes
related talks
27mShipping a SaaS backend using Vapor
27mFull Stack Development with Swift and Vapor
36mFull stack development with Swift and Vapor
21mThe Adventures and Misadventures of Building a Trivia app with Vapor
19mServer-side Swift Using Vapor
41mGetting started with Server-Side Swift and Vapor
31mFrom iOS- to Vapor developer
17mThe Success Story of Server-Side Swift at Cultured Code