Tail latency and fan-out
On This Lesson 4 topics
This lesson is planned. The outline below lists the topics it will cover. The full lesson has not been published yet.
In this lesson
see how tails amplify across calls.
Planned topics
- If a request calls N services in parallel, it waits for the slowest.
- With 10 calls each slow 1% of the time, about 10% of requests see a slow call.
- Mitigations: fewer calls, timeouts, hedged requests, caching, removing calls from the critical path.
- Measure percentiles per hop.
Draft outline details
Demonstration
the calculation 1 - 0.99^10, about 9.6%, and a trace where one slow hop dominates.
Check your understanding
What happens to tail latency as fan-out grows?
Show answer
The chance of hitting a slow call rises, so overall tail gets worse.
Practice
Apply this lesson with the module's lab, exercises and quiz.
Marking a lesson complete is optional and you can undo it any time. Your progress stays in this browser.