We don't want requests clogging up an already stressed system. Optimally, this won't be an issue, requests should go through in milliseconds - however if something is wrong with the network and it becomes latent, we don't want connections backing up in the pipeline awaiting responses.
A timeout approximately 50% of the update interval would suffice.
marc-costello
commented
7 years ago
We don't want requests clogging up an already stressed system. Optimally, this won't be an issue, requests should go through in milliseconds - however if something is wrong with the network and it becomes latent, we don't want connections backing up in the pipeline awaiting responses.
A timeout approximately 50% of the update interval would suffice.