Expand description
Dynamic thread pool for blocking operations
The pool would lazily generate thread according to the workload and spawn up to a total amount
of you machine’s logical CPU cores * 5
threads. Any spawned threads kept idle for 30 seconds
would be recycled and de spawned.
*. Settings are configuable through env variables.
§Example:
use std::env::set_var;
#[actix_rt::main]
async fn main() {
// Optional: Set the max thread count for the blocking pool.
set_var("ACTIX_THREADPOOL", "30");
// Optional: Set the min thread count for the blocking pool.
set_var("ACTIX_THREADPOOL_MIN", "1");
// Optional: Set the timeout duration IN SECONDS for the blocking pool's idle threads.
set_var("ACTIX_THREADPOOL_TIMEOUT", "30");
let future = actix_dynamic_threadpool::run(|| {
/* Some blocking code with a Result<T, E> as return type */
Ok::<usize, ()>(1usize)
});
/*
We can await on this blocking code and NOT block our runtime.
When we waiting our actix runtime can switch to other async tasks.
*/
let result: Result<usize, actix_dynamic_threadpool::BlockingError<()>> = future.await;
assert_eq!(1usize, result.unwrap())
}
Structs§
- CpuFuture
- Blocking operation completion future. It resolves with results of blocking function execution.
Enums§
- Blocking
Error - Blocking operation execution error
Functions§
- run
- Execute blocking function on a thread pool, returns future that resolves to result of the function execution.