Skip to content

Load Balancing

Distribute incoming work across multiple viable receivers by capacity, health, or policy so no part is overloaded while usable capacity sits idle.

The Diagnostic Story

Symptom: Some receivers are saturated while others sit idle, and the system behaves as if capacity is scarce even though total capacity may be adequate. One server is overloaded while five others handle almost nothing. One team absorbs every request because it is the most visible. One region, warehouse, or service desk becomes a hot spot while equivalent alternatives are underused.

Pivot: Change how work is assigned. Route, distribute, or redirect incoming flow across multiple viable receivers according to a policy that accounts for capacity, health, load, locality, or priority. The key move is not simply splitting evenly; it is sending work where it can actually be handled.

Resolution: Localized overload declines as idle capacity is put to use and the distribution policy adapts to receiver health and load. The balancing layer does not drop work, and unhealthy receivers are removed or downweighted. Total demand is unchanged, but poor assignment is no longer making the system behave as if capacity were insufficient.

Reach for this when you hear…

[platform SRE] “Two of our six nodes are at one hundred percent and the other four are doing nothing because the load balancer is still sending everything to the first healthy response.”

[hospital intake] “The ER is on diversion while the urgent care centers nearby have open beds, so we need to route patients differently rather than just expanding the ER.”

[workforce management] “Half the team is drowning in tickets while the other half has been sitting at sixty percent capacity for weeks because we never looked at the assignment queue.”

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
  • Capacity Signal
  • Distribution Policy
  • Feedback: Outputs influence inputs.
  • Health Check
  • Monitoring: Continuously observing a system's state to detect deviation from expected behavior and trigger a response, separating genuine signal from routine noise.
  • Resource Pooling
  • Routing Rule

Also references 13 related abstractions

  • Constraint: Limits possibilities to guide outcomes.
  • Coupling: Interdependence among subsystems.
  • Fairness: Judging whether an allocation or procedure treats comparable parties impartially according to a defensible standard, given that multiple such standards can conflict.
  • Flow: Structured movement of energy, matter, or information.
  • Network: Models interactions between components.
  • Observability: Infer internal state externally.
  • Queueing: Organizes tasks into a waiting line based on arrival and service rates.
  • Resilience: Absorb shocks and adapt.
  • Resource Management: Allocation of finite assets.
  • Robustness: Maintain functionality under stress.

Variants

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

Capacity-Aware Routing · implementation variant · recognized

Route work or flow based on current or expected capacity across multiple receiving paths.