Prerequisites
What it is
Distributed tracing records the path of a request through multiple services as a tree of spans.
Why it matters
Modern requests often cross gateways, services, caches, queues, and databases. Tracing shows where time is spent and where failures start.
How it works
A trace ID is created at the edge and propagated through headers or message metadata. Each service creates spans for its work and exports timing, attributes, and errors.
When to use it
Use tracing for microservices, APIs with latency SLOs, complex fan-out, background workflows, and third-party integrations.
Tradeoffs / failure modes
Sampling can hide rare issues, missing propagation breaks traces, and high-cardinality attributes can become expensive.
Tiny example
A slow checkout trace shows 30ms in the API, 80ms in inventory, and 1.8s waiting on a payment provider.
Related guides and systems
See Webhooks, Job Queues, API Versioning, Netflix, Slack, ChatGPT, and GitHub.
Seen in deeper dives
Capability
API Versioning
Backward-compatible API evolution with URL, header, and content negotiation strategies
Capability
Job Queues & Background Workers
Async task processing with retries, dead-letter queues, and priority scheduling
Capability
Webhooks
Reliable event delivery to external systems with retry, HMAC verification, and monitoring
Case Study
ChatGPT
Serving large language model inference to hundreds of millions of users
Case Study
GitHub
The world's largest code hosting platform powered by Git at scale
Case Study
Netflix
Global video streaming with personalized content delivery
Case Study
Slack
Channel-based messaging for the enterprise workplace