Skip to content

Load Shedding

Deliberately drop, deny, or defer lower-priority load under overload so critical function stays within viable bounds.

The Diagnostic Story

Symptom: Total demand has exceeded what the system can safely handle, and attempting to serve everything is degrading or about to collapse critical function. The queue is growing without bound. Essential paths are waiting behind nonessential ones. The system is treating all load as equally serviceable even though some of it is far more expendable than others.

Pivot: Make sacrifice explicit. Classify load by criticality, urgency, or recoverability, then deliberately drop, deny, defer, or deprioritize the lower-priority portion. The policy names what is shed, under what conditions, and how the shed load is handled, rather than letting collapse choose arbitrarily.

Resolution: Critical function remains viable because it retains capacity and queue growth is bounded. Shed load is handled predictably: rejected, expired, deferred, or rerouted with acknowledgment. The sacrifice is bounded by explicit fairness and safety rules rather than by whoever happens to fail first.

Reach for this when you hear…

[grid operator] “We drop residential blocks in rotation before the frequency deviation gets bad enough to trip generation, because losing a block is recoverable and losing a generator is not.”

[emergency triage nurse] “When the waiting room has forty people and I have two doctors, I have to triage or the critically ill patients die waiting behind the sprained ankles.”

[platform SRE] “During the incident we started rejecting recommendation requests and analytics writes so the checkout path could stay up.”

Abstractions this archetype builds on — directly (a source ingredient) or as a related pattern. Links follow the typed catalog namespace.

Built directly on (8)

  • Admission Control
  • Boundary: Defines system limits.
  • Graceful Degradation
  • Observability: Infer internal state externally.
  • Prioritization: Ordering competing claims on finite resources by a value or urgency metric to produce a ranked sequence of action under constraint, making explicit what gets done first and what does not get done at all.
  • Resource Management: Allocation of finite assets.
  • Selective Rejection
  • Threshold: Safe vs harmful levels.

Also references 9 related abstractions

  • Constraint: Limits possibilities to guide outcomes.
  • Coupling: Interdependence among subsystems.
  • Fail-Safe: Default to safe state on failure.
  • Fault Tolerance: Continue operating under failure.
  • Flow: Structured movement of energy, matter, or information.
  • Queueing: Organizes tasks into a waiting line based on arrival and service rates.
  • Resilience: Absorb shocks and adapt.
  • Scheduling: Organizing tasks over time.
  • Trade-offs: Balancing competing priorities.

Variants

Narrower or domain-specific specializations that share this archetype's core structure. Recognized variants are established; candidate variants are provisional.

Selective Rejection · risk or failure variant · recognized

Reject or drop lower-priority inflow to preserve viability for remaining work or users.