I think there would be value in adding a flag that displays the WASI hostcalls made as part of running a function. This could be used to as part of investigating why a function's latency has unexpectedly increased.
The wasi-common crate used by wasmtime has trace level log statements that contain the hostcalls made as well as their arguments and return values. This crate could expose a way to display those log messages on the console.
For example, when running wasmtime directly, I can set RUST_LOG=wasi_common=trace to see details about the WASI hostcalls.
I think there would be value in adding a flag that displays the WASI hostcalls made as part of running a function. This could be used to as part of investigating why a function's latency has unexpectedly increased.
The
wasi-commoncrate used bywasmtimehastracelevel log statements that contain the hostcalls made as well as their arguments and return values. This crate could expose a way to display those log messages on the console.For example, when running
wasmtimedirectly, I can setRUST_LOG=wasi_common=traceto see details about the WASI hostcalls.