docs: Update readme, scope has changed

This will no longer attempt to copy what checkmk_kube_agent does, it
will experiment with new ways of doing things instead.

Author: relrod

README.md

@@ -1,52 +1,7 @@
 # cmk-rustik
 
-A currently-unofficial re-implementation of
-[checkmk_kube_agent](https://github.com/Checkmk/checkmk_kube_agent/) in Rust.
-
-**Currently a work-in-progress/experiment.**
-
-Right now a focus is matching the Python implementation pretty-darn-close to
-1:1. This means: not modifying the "wire protocols" (i.e. the JSON schemas)
-sent and expected.
-
-This makes each individual component drop-in.
-
-## Components
-
-#### `metrics-cache` _(aka: `cluster_collector` in the Python original)_
-
-This is effectively an in-memory cache server that functions over HTTP and
-authenticates requests against the Kubernetes cluster in which it is running.
-
-Because the cache is embedded (not some external service), **one instance**
-is expected to be running per cluster. (In the future, we could allow for an
-external backing cache or some kind of tiered L1/L2 layered caching and then
-permit multiple instances of `metrics-cache` to run for HA.)
-
-Requests are expected to have a ServiceAccount token as a Bearer token. This is
-validated by Kubernetes (internally, a `TokenReview` is sent to Kubernetes and
-the result is validated and then checked against the allowlist which is
-configured by `metrics-cache`'s CLI arguments).
-
-There are/will be three internally-maintained caches in total:
-
-* Container metrics (for metrics coming from cAdvisor)
-* Machine sections (for storing a copy of the `check_mk_agent` output to gather
-  *node* information)
-* `metrics-fetcher` metadata (for storing metadata about the `metics-fetcher`
-  instances sending data to the `metrics-cache`).
-
-#### `metrics-fetcher` _(aka: `node_collector` in the Python original)_
-
-A "metrics fetcher", as the name suggests, fetches metrics from some source. It
-then forwards those to the metrics cache instance for consumption by Checkmk.
-
-The metrics fetcher is one binary that can be run in two different modes:
-
-* "cAdvisor mode" - fetch container metrics from a cAdvisor instance
-* "cmk-agent mode" - run and collect sections from a `check_mk_agent` script
-
+WIP Kubernetes experiments with Checkmk.
 
 ## License
 
-GPLv2, same as the Python original.
+GPLv2