reqwest offers a read_timeout option, which can be set to avoid the process getting stuck indefinitely while still allowing large downloads to take a long time. I'm trying to figure out what the correct value for a widely deployed CLI application is that often runs in unreliable environments such as GitHub Action. For that I'm wondering, what is the unit of a read, what counts as read? For example, if a read timeout of 10s failed, does this mean that no package was received from the server in the last 10s? Or does it require a specific throughput, e.g. a certain number of KB?
In that context, do we need to set a connect_timeout if we set a read_timeout, or should we set them to separate values? The documentation doesn't really say anything about best practices using these functions.
reqwest offers a
read_timeoutoption, which can be set to avoid the process getting stuck indefinitely while still allowing large downloads to take a long time. I'm trying to figure out what the correct value for a widely deployed CLI application is that often runs in unreliable environments such as GitHub Action. For that I'm wondering, what is the unit of a read, what counts as read? For example, if a read timeout of 10s failed, does this mean that no package was received from the server in the last 10s? Or does it require a specific throughput, e.g. a certain number of KB?In that context, do we need to set a
connect_timeoutif we set aread_timeout, or should we set them to separate values? The documentation doesn't really say anything about best practices using these functions.