Add WatchStreamExt::default_backoff shorthand

[clux]

This change adds a simplified recommended backoff construction for watcher streams:

-let stream = watcher(api, config).backoff(watcher::default_backoff()).
+let stream =  watcher(api, config).default_backoff();

This is advantageous because it's hard to advocate that users use backoff (fundamentally a good behaviour action) when the usage is verbose. So this makes it less verbose.

To do this properly we have made a new DefaultBackoff struct that encapsulates the type. This is slightly more verbose than relying on impl Backoff from watcher::default_backoff, but it has the benefit of working in the stream setting where we are limited by existential type constraints.

This also deprecates the watcher::default_backoff fn due to it now generally being more easily invoked elsewhere, and by the explicit type. This could be debated.

old, now irrelevant commentary, left in a collapsible

The slight downside here is that `watcher::default_backoff()` returns an opaque type and cannot be used in the shortcut stream ext method so have exposed just our `ExponentialBackoff` default as an internal fn (which cannot be const), along with a const reset timeout and re-implemented the function in `WatchStreamExt::default_backoff` referencing our parameters.

An alternative could be doing the defaulting in the generally opaque type that most users don't interact with:

// convenience for watchstreamext users
impl Default for ResetTimerBackoff<backoff::ExponentialBackoff> {
    fn default() -> Self {
        use crate::watcher::{default_exponential_backoff, DEFAULT_RESET_DURATION};
        let bo = default_exponential_backoff();
        Self::new(bo, DEFAULT_RESET_DURATION)
    }
}

but i think that doesn't really help much. we still end up with an implementation in WatchStreamExt that is almost as big (except without the internal imports) and it's an awkward inversion of utils depending on core stuff. I'd probably rather keep the backoff parameters somewhere prominent.

[codecov]

Codecov Report

Merging #1232 (7bbf028) into main (a6a627c) will decrease coverage by 0.13%.
The diff coverage is 0.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1232      +/-   ##
==========================================
- Coverage   72.06%   71.94%   -0.13%     
==========================================
  Files          71       71              
  Lines        5714     5725      +11     
==========================================
+ Hits         4118     4119       +1     
- Misses       1596     1606      +10     

Impacted Files	Coverage Δ
kube-runtime/src/controller/mod.rs	34.00% <0.00%> (ø)
kube-runtime/src/utils/watch_ext.rs	0.00% <0.00%> (ø)
kube-runtime/src/watcher.rs	39.31% <0.00%> (-2.92%)	⬇️

... and 1 file with indirect coverage changes

[Dav1dde]

Existential types would be nice ...

I like the helper, especially since this is already the controller default behaviour.

Have you considered just encapsulating the default backoff in a new-type?

pub struct DefaultBackoff(InnerStrategyNotPublic)

impl Default for DefaultBackoff {
    fn default() -> Self {
        Self(default_exponential_backoff())
    }
}

impl Backoff for DefaultBackoff {
    fn next_backoff(&mut self) -> Option<Duration> {
        self.0.next_backoff()
    }
}

And then just expose DefaultBackoff in all the APIs (including the default_backoff()) function.

[clux]

Have you considered just encapsulating the default backoff in a new-type?

pub struct DefaultBackoff(InnerStrategyNotPublic)

And then just expose DefaultBackoff in all the APIs (including the default_backoff()) function.

That suggestion does have legs. I might change the naming a bit because you'd end up with impl Default for DefaultBackoff and you'd construct it with DefaultBackoff::default() which feels awkward. Perhaps watcher::Backoff::default() and then deprecating the old default_backoff fn will be helpful here.

EDIT: no, watcher::Backoff is probably a bad name given there are so many different Backoff suffixed things. 🤔

[clux]

Ok, have done this properly now. It is still somewhat hampered by the fact that ResetTimerBackoff is actually public, but have at least managed to keep that type out of DefaultBackoff and the WatchStreamExt!

[clux]

Giving a ping to @nightkr here in case of extra opinions since she wasn't cc'd on originally.

[nightkr]

LGTM