Requests to URLs that do not have any content registered with the client anymore still get their requests forwarded to the client, due to the way URL generation is designed. Currently empty responses to URLs that are not registered anymore are not cached by the server. An outside actor could use a single URL like this to make the client disconnect and reconnect to the server, simply by sending a few too many requests to that URL in a short time period.
It would be good if it was a cache's requirement to cache responses to invalidated URLs. This should not take up much storage in the cache, as it's simply a cached 404 response without any body.
Requests to URLs that do not have any content registered with the client anymore still get their requests forwarded to the client, due to the way URL generation is designed. Currently empty responses to URLs that are not registered anymore are not cached by the server. An outside actor could use a single URL like this to make the client disconnect and reconnect to the server, simply by sending a few too many requests to that URL in a short time period.
It would be good if it was a cache's requirement to cache responses to invalidated URLs. This should not take up much storage in the cache, as it's simply a cached 404 response without any body.