Course contentsShow
← System Design
Lesson 1108 of 1,91928. Timeouts and DeadlinesFree lesson

What is Deadline Propagation

Understanding how timeouts cascade through distributed call chains and why explicit deadline passing is critical.

What is Deadline Propagation

You'll learn the difference between timeouts and deadlines, and why propagating absolute deadlines prevents cascading timeout failures.

The Difference Between Timeouts and Deadlines

A timeout is a relative duration: "wait up to 5 seconds." A deadline is an absolute timestamp: "must complete by 14:32:15.500 UTC."

When Service A calls Service B with a 5-second timeout, and B calls C with another 5-second timeout, you've accidentally given the entire chain 10 seconds. Each hop resets the clock. This is the timeout chain problem you've already learned.

Deadline propagation solves this by passing an absolute deadline down the call chain instead of durations. Service A sets a deadline (e.g., "now + 5 seconds") and passes that exact timestamp to B. When B calls C, it forwards the same deadline, not a new timeout.

How It Works in Practice

Request arrives at Service A at T=0
A sets deadline: T=5000ms (absolute)
A→B: "Complete by T=5000ms"
B→C: "Complete by T=5000ms" (same deadline)
C checks: only 3000ms remain, uses that

Every service checks how much time is left against the original deadline. No service can accidentally extend the total time budget.

Why This Matters

Without deadline propagation, deep call chains fail slowly—each hop times out independently, wasting resources at every level. With deadlines, services near the end of the chain see "0ms remaining" and immediately fail fast, preventing wasted work.

Deadline propagation is timeout propagation done right: instead of passing shrinking durations (error-prone), you pass a single absolute timestamp.

Key Takeaway: Deadline propagation uses absolute timestamps instead of relative timeouts to ensure the entire distributed call chain respects a single, consistent time budget.