Glossary · Networking

Load Balancer

A load balancer distributes incoming traffic across multiple servers, using health checks to route around failures and improve availability and capacity.

Updated October 5, 2026

A load balancer distributes incoming connections or requests across a pool of backend servers. It works at layer 4 (TCP/UDP, by address and port) or layer 7 (HTTP, by host, path or header) and removes unhealthy servers from rotation.

Why it matters

A single server is a single point of failure. With a load balancer you can add capacity, deploy by draining one node at a time and survive a server crash without downtime. The load balancer itself must then be redundant, or it becomes the new single point of failure.

Example

Common algorithms:

  • Round robin: each server in turn. Simple, ignores load.
  • Least connections: picks the server with the fewest active connections. Better for long requests.
  • IP hash / cookie affinity: keeps a client on the same server ("sticky sessions").

An HAProxy backend with health checks:

backend web
    balance leastconn
    option httpchk GET /healthz
    server app1 10.0.0.11:8080 check inter 5s fall 3 rise 2
    server app2 10.0.0.12:8080 check inter 5s fall 3 rise 2

Three failed checks (fall 3) remove a server; two passes (rise 2) restore it.

Common mistakes

  • Health checks that test only the port, so a server with a broken database stays in rotation.
  • Relying on sticky sessions instead of shared session storage, which loses sessions when a node dies.
  • Not testing failover, so capacity after losing one node is insufficient.
  • Monitoring only the load balancer address, which hides a single unhealthy backend.

Get notified when this term updates

Plus occasional practical tips on uptime and monitoring, nothing else. Unsubscribe anytime.

Keep reading

Guide · Hosting & CDNVPS vs Shared Hosting: Which Is More Reliable for Your Website?Compare shared hosting and VPS on reliability: resource isolation, noisy neighbors, failure modes, who maintains what, and when moving to a VPS actually reduces downtime.7 min readGuide · WordPressWordPress "Error Establishing a Database Connection": How to Fix ItFix the WordPress error establishing a database connection: check wp-config.php credentials, test MySQL, run repair mode with WP_ALLOW_REPAIR and spot host limits.8 min readGuide · WordPressWordPress "Allowed Memory Size Exhausted" Error: How to Raise the LimitFix the WordPress "Allowed memory size exhausted" fatal error by raising WP_MEMORY_LIMIT, php.ini, .user.ini or .htaccess limits and finding what uses the memory.8 min readError fix · Server errorsHTTP 508 Loop Detected and Resource Limit Is Reached: FixesFix HTTP 508 Loop Detected. It means a request cycle on WebDAV, and on CloudLinux hosts it appears as Resource Limit Is Reached. Find the cause and fix it.4 min readError fix · Developer errorsNode.js ENOMEM "Cannot Allocate Memory": Find What Ran OutFix Node.js spawn ENOMEM and "Cannot allocate memory" errors: tell OS memory exhaustion from heap limits, check the OOM killer, swap and container limits.4 min readGlossary · MonitoringBlack-Box MonitoringBlack-box monitoring tests a system from the outside, the way a user would, to detect symptoms like errors and slowness without inspecting internals.

This page is general information written by our editorial team and updated when documentation changes. Servers and providers differ, so test changes safely first. See our editorial policy and disclaimer.