Bigger sites are more likely to use microservices and multiple failovers so individual components can be upgraded without downtime. Smaller sites likely have a single point of failure that needs to go down to make changes (e.g. 1 single master database that needs to be write locked for a schema change).