CASE STUDY
Event-driven messaging over Kafka and Amazon SQS, with service discovery, an API gateway and circuit breakers holding 99.99% uptime.
Role
Full Stack Developer — service design and messaging
Project
QSS Technosoft · Sep 2022 – present
Containerised microservices
9
Requests handled
100K+ / day
Uptime
99.99%
Throughput gain
40%
Latency reduction
60%
CONTEXT
A multi-tenant SaaS platform where a single user action fans out into work that other parts of the system care about — check-ins, queue position updates, notifications and calendar sync. Each organisation's data has to stay separate on shared infrastructure.
THE PROBLEM
Calling every interested service synchronously made the request path as slow as its slowest dependency and as fragile as its least reliable one. Adding a new consumer of an existing event meant editing and redeploying the producer.
HOW IT WORKS
Producers publish once; consumers subscribe independently. A circuit breaker keeps one failing downstream from taking the request path with it.
TRADE-OFFS
What was chosen, what it was chosen over, and why.
ConsideredSynchronous REST calls between services
WhyA synchronous chain makes every caller wait on every callee and couples deploys together. Publishing an event lets a new consumer subscribe without the producer knowing it exists, and keeps the user-facing request fast.
ConsideredOne broker for everything
WhyThey answer different questions. A retained, replayable log suits events several services consume and may need to reprocess; a managed queue suits point-to-point work that just needs to be done once, without the operational overhead.
ConsideredRetries and timeouts alone
WhyRetries against an already-failing dependency add load to the thing that is falling over. Breaking the circuit fails fast and lets the rest of the platform stay inside its uptime budget while the dependency recovers.
ConsideredHard-coded service addresses per environment
WhyInstances move, scale and get replaced. One entry point plus discovery keeps routing, authentication and TLS termination in a single place instead of duplicated across nine services.
ConsideredPer-endpoint permission checks in controllers
WhyMulti-tenant separation enforced per endpoint is one missed check away from a cross-tenant leak. Putting the check in the security layer with JWT, MFA and OAuth 2.0 makes the safe path the default.
IN HINDSIGHT