The webservers use rate limiting per IP address. It reduces the risk of overload from certain types of attack (so-called DoS attacks) and somewhat badly programmed websites.
Only extreme cases are blocked. When a request is blocked, a HTTP error is provided, either "403 Forbidden" or "503 Service Unavailable".
The limits are handled differently for Web Starter than for greater web hosting packages.
Web Starter
Web Starter uses the Apache module mod_evasive.
The limits per IP address are:
Max page views started per 2 seconds: 10
Max page elements per second: 120
Lock-out period: 10 seconds
Web Medium and greater
Web Medium and greater use Nginx as a reverse proxy and for handling rate limiting.
Innstilling | Default | Burst |
POST requests/sec |
10 |
40 |
POST requests/min |
50 |
200 |
All requests/sec |
100 |
200 |
All requests/min |
1000 |
2000 |
HTTP 1 connections |
600 |
n/a |
HTTP 2 connections |
6000 |
n/a |
Technically inclined customers may read about how rate limits in Nginx work: