Course contentsShow
← System Design
Lesson 1123 of 1,91928. Timeouts and DeadlinesFree lesson

Client-Side vs Server-Side Timeout Enforcement

Understand when each side should enforce timeouts and how to coordinate both approaches.

Client-Side vs Server-Side Timeout Enforcement

What you'll learn: How to decide which side of a request should enforce timeouts and why coordinating both matters for reliable distributed systems.

The Dual Enforcement Problem

Both clients and servers can enforce timeouts, but they serve different purposes. Client-side timeouts protect the caller from waiting indefinitely—they're about resource management and user experience. Server-side timeouts protect backend resources from runaway operations and ensure fair resource allocation among all callers.

When Each Side Should Enforce

Client-side enforcement is essential when:

  • You need to prevent thread/connection pool exhaustion
  • User-facing latency budgets must be met
  • You're aggregating multiple downstream calls and need to abort early
  • The client has tighter SLAs than the server's natural response time

Server-side enforcement is critical when:

  • Long-running database queries or computation could monopolize resources
  • You need to prevent abandoned work (client disconnected but server keeps processing)
  • Resource cleanup requires explicit cancellation
  • You want consistent timeout behavior across all clients

Coordinating Both Approaches

The ideal setup uses both. The client sets an aggressive timeout (say, 2 seconds) and propagates a deadline to the server. The server enforces its own timeout (perhaps 2.5 seconds as a backstop) and checks the propagated deadline to abort work early if the client's budget is already exhausted.

Without coordination, you get waste: clients timeout and retry while servers continue processing abandoned requests, amplifying load. With coordination, servers can stop work immediately when deadlines expire, freeing resources for productive requests.

Key Takeaway: Client-side timeouts protect the caller and meet user SLAs; server-side timeouts protect backend resources and prevent wasted work. Deploy both and coordinate them through deadline propagation for optimal reliability and efficiency.