For the current version of node-fetch that the client uses, the default timeout is 0 (OS-dependent). We're noticing our service that only makes calls to fusion auth go to 100% CPU usage at times and can't tell if it's due to requests getting stuck in flight because of the lack of a timeout.
For systems like Linux, the timeout is usually 3 minutes, which is way too long IMO.
https://stackoverflow.com/questions/1045964/how-to-view-change-socket-connection-timeout-on-linux
https://github.com/node-fetch/node-fetch/tree/2.x?tab=readme-ov-file#options
For the current version of
node-fetchthat the client uses, the default timeout is 0 (OS-dependent). We're noticing our service that only makes calls to fusion auth go to 100% CPU usage at times and can't tell if it's due to requests getting stuck in flight because of the lack of a timeout.For systems like Linux, the timeout is usually 3 minutes, which is way too long IMO.
https://stackoverflow.com/questions/1045964/how-to-view-change-socket-connection-timeout-on-linux
https://github.com/node-fetch/node-fetch/tree/2.x?tab=readme-ov-file#options