RateLimitPolicy does not work after being modified

[trepel]

Overview

It seems that after RateLimitPolicy (RLP) CR is modified - e.g. limit is changed to some other number - it stops working. By that I mean that the 429 Too Many Request is never returned, one receives as many 200 OKs as they want.

The only suspicious entry from log is found in Limitador pod log. The following entry appears in that log after each hit of API endpoint that is rate-limited by modified RLP:

thread 'main' panicked at limitador/src/storage/in_memory.rs:106:58:                                      
called `Option::unwrap()` on a `None` value                                                               

Once limit it reverted back to original value everything starts working as expected.

This has been discovered using Kuadrant nightly build from November 20

Steps to reproduce

Create 1 GW, 1 HTTPRoute and one RLP attached to GW (does not matter, could be also attached to HTTPRoute instead). Make sure everything works as expected and the 429 is returned after expected no. of 200s.
Edit the RLP so that different no. of requests is allowed.
Try again, this time 429 is not returned as expected, one continues getting 200s. And see Limitador pod logs for the entry mentioned above.

[alexsnaps]

Is this happening with 1.6 as well?

[alexsnaps]

Seems to be a v2.0.0-dev only issue, probably (partially?) related to #344 around these changes: 0c08ce4