Bloqueo de IP
Un bloqueo aplicado a una dirección IP concreta que le impide acceder a un sitio web o servicio.
Un bloqueo de IP ocurre cuando un sitio web bloquea todas las solicitudes de una dirección IP concreta. Desde la IP bloqueada, el sitio devuelve un error (403, 429 o una página de bloqueo personalizada) en lugar del contenido solicitado.
Los sitios bloquean IPs por varias razones: demasiadas solicitudes en poco tiempo (rate limiting), patrones de solicitud que parecen automatizados (tiempos constantes, cabeceras idénticas), acceso a endpoints restringidos o coincidencia con un rango conocido de IP de centro de datos o maliciosas.
Cómo detectar un bloqueo: si tus solicitudes empiezan a devolver errores 403, CAPTCHA o respuestas vacías tras funcionar correctamente, la IP puede estar bloqueada. Prueba la misma solicitud desde otra IP para confirmarlo.
Los bloqueos temporales expiran tras un periodo de enfriamiento -de minutos a horas. Los bloqueos permanentes bloquean la IP indefinidamente hasta que se revise manualmente o el ISP recicle la dirección.
Los proxies rotativos evitan los bloqueos de IP repartiendo las solicitudes entre muchas IPs. Ninguna IP acumula suficientes solicitudes para alcanzar el umbral de bloqueo. Al usar pools residenciales con millones de IPs, la probabilidad de toparte con una IP previamente bloqueada es muy baja.
Los proxies no evitan todos los bloqueos: los sitios también bloquean por user-agent, patrones de cookies y huella del navegador. La rotación de IP es una capa; combínala con patrones de solicitud realistas y user-agents variados para obtener mejores resultados.
What a ban actually attaches to
An IP ban refuses traffic from an address. What varies, and what decides your response, is the scope: a single address, a subnet, an entire autonomous system, or a rule that fires on the combination of address and something else about your request.
Duration varies just as much. Some bans expire in minutes, some persist for weeks, and some are not bans at all but rate limits that look identical from outside. A 403 on every request from one address across days is a ban. A 429 that clears after ten minutes is a limit.
Sites also ban softly. Instead of refusing you they serve reduced content: default prices, an empty catalogue, stale inventory. This is worse than a block because your scraper reports success and your data is quietly wrong.
Reading the response before reacting
The status code tells you which problem you have, and each one has a different fix:
Diagnosis
403 on the first request -> address class rejected, change proxy type
403 after many requests -> address burnt, rotate or rest it
429 -> rate limit, spread across more addresses
200 with suspiciously thin data -> soft ban, compare against a clean fetch- Compare a suspicious response against one fetched through a different address type. Soft bans only show up in comparison.
- If a fresh static address arrives already blocked, that is inherited history and we will replace it.
- Resting an address often works better than rotating fast. Many limits decay with time rather than with traffic volume.
- Never retry a hard 403 in a loop. It costs traffic on metered products and confirms the pattern that got you blocked.
Misconceptions about bans
A new address does not clear an account flag
If the account is flagged, the address was only part of the evidence. Moving it to a clean IP does not undo the record.
Rotating faster is not a fix for rate limits
The counter is per address. Spreading across more addresses is the fix; cycling one address faster is not.
Not every block is about you
Shared and previously used addresses carry other people's history, which is what dedicated addresses are for.
A ban is not always visible
Soft bans return 200. Treat unexpectedly clean or thin data as a symptom.
Términos relacionados
Ver esto en práctica
¿Listo para usar bloqueo de ip?
SotaProxy te da acceso a proxies residenciales rotativos, móviles, de centro de datos e ISP. Sin compromiso mínimo.
Empezar