Eliza Weisman

IMO using two counters rather than a gauge is definitely more correct for these metrics, so I'm 👍 on this change.

> If we deprecate, I would doc hide it. > > > > I don't *love* the names. What about: > > > > * `from_blocking_std` > > * `from_nonblocking_std`...

Rather than a `mpsc` channel, I would recommend using a [`tokio::sync::Semaphore`](https://docs.rs/tokio/latest/tokio/sync/struct.Semaphore.html). The task awaiting the countdown should call [`Semaphore::acquire_many`](https://docs.rs/tokio/latest/tokio/sync/struct.Semaphore.html#method.acquire_many) with the initial value of the countdown latch, while other tasks...

I can open a PR to add the example to the docs.

I think that we'd probably prefer to have something like this in `tokio_util`, rather than the core `tokio` crate, just because it's possible to implement it entirely outside of `tokio`,...

> I think one solution is to define new `feed` and `send` methods that take a `&mut Option`, and the caller is expected to pass in a `Some`. Then the...

Thank you for working on this! The current queue sizes and other parameters were chosen pretty arbitrarily in the interest of "just getting something working", so I had been hoping...

The reason that the retention periods for `console-subscriber` and for the `tokio-console` CLI are different is because they configure very different things. The `console-subscriber` retention period configures how long the...

> Would it be possible to only keep the objects that are a parent? I haven't looked yet, but if you think it can be done I could try working...

It looks like this change requires a MSRV bump, as the new `ratatui` seems not to support Rust 1.64.0: https://github.com/tokio-rs/console/actions/runs/7763271246/job/21175134601#step:8:1