Load balancer
Infrastructure that spreads incoming requests across many backend servers, a form of reverse proxy you meet from the outside when scraping.
A load balancer distributes incoming traffic across a pool of backend servers so no single machine is overwhelmed and the service survives any one of them failing. It is a specific job within the reverse-proxy family: sitting in front of a site's own infrastructure, fielding requests, and handing each to whichever backend is healthy and least busy. If a site stays up while individual servers are rebooted behind it, a load balancer is usually why.
You meet it from the outside, which is what makes it worth knowing rather than just server-side trivia. When you connect to a large site you are almost always talking to its load balancer, not to an origin machine, and the balancer commonly carries the same edge functions a reverse proxy does: TLS termination, health checks, and often the rate limiting and bot rules that decide whether your request proceeds. A refusal can originate here, before any application server is involved.
It also explains a behaviour that confuses scrapers: apparent inconsistency between identical requests. Because successive requests can be routed to different backends with slightly different state or cache contents, the same URL can return marginally different responses moments apart, and a session that assumes it is always talking to one machine can be surprised. This is distinct from deliberate variation; it is just the pool showing through.
The term overlaps heavily with reverse proxy and CDN, and the honest distinction is one of emphasis. A reverse proxy is the general category; load balancing is one function it performs, alongside caching and filtering; a CDN is that pattern operated at geographic scale. A dedicated load balancer may do only the distribution, while most reverse proxies and every CDN balance load as part of a larger job. For a scraper the practical takeaway is the same: the thing answering you is infrastructure in front of the site, and it, not the site, often decides your fate.
Frequently asked questions
Is a load balancer the same as a reverse proxy?
Load balancing is one of the jobs a reverse proxy does, so every load balancer is reverse-proxy-like but not every reverse proxy is only a load balancer. A reverse proxy also caches, terminates TLS and filters traffic; a dedicated load balancer may just distribute requests across backends. They overlap enough that the same software often fills both roles.
Why does a site behind a load balancer give me inconsistent results?
Because successive requests can land on different backend servers whose cache or state differs slightly, so identical requests return marginally different responses. It is not deliberate variation aimed at you; it is the server pool showing through. Code that assumes it is always talking to one machine can misread this as a bug or a block.
Can a load balancer block my scraper?
Often it is what does. Sitting in front of the origin, a load balancer commonly enforces the same edge functions a reverse proxy does, including rate limiting and bot rules, so a refusal can happen there before any application server sees your request. That is also why the site's own logs may show nothing of a request that was rejected at the balancer.
Back to the full glossary.