Docs regarding proxy-cpu-limits seem misleading

Hi!

The docs regarding the config.linkerd.io/proxy-cpu-limit annotation seem a bit misleading to me.
https://github.com/linkerd/website/blob/main/linkerd.io/content/2.12/tasks/configuring-proxy-concurrency.md

There it is explicitly stated that:

> Unlike Kubernetes CPU limits and requests, which can be expressed in milliCPUs, the proxy-cpu-limit annotation should be expressed in whole numbers of CPU cores. Fractional values will be rounded up to the nearest whole number.

I think the last sentence can be removed or rephrased, as specifying e.g. 0.5 as the limit translates perfectly fine into a cpu.limit of "500m" on the containers.
However, I'm not sure if this is intended and documented wrong, if I misunderstood the docs or if it is a bug 😉

Review linkerd.io/content/2.12/tasks/configuring-proxy-concurrency.md and the documentation for the config.linkerd.io/proxy-cpu-limit annotation. First determine whether fractional CPU values are intended to work as described or are rounded up, then update the quoted wording accordingly. Done means the page accurately explains fractional values and their resulting container limits.

Written by the indexing model from the issue text.