[CAE-342] Initial sentinel issues fix

[htemelski]

Description

Fixed a couple of issues with the sentinel implementation

1. Indefinite reconnects without taking into consideration maxCommandRediscovers
2. Client losing it's watch status if there are multiple topology changes during a transaction

Fixed an issue with the testing setup itself, using the new docker image and setting the requirepass and masterauth configuration settings in a persistent way.

Enabled the majority of sentinel tests, added new test for fix 1.
Still skipping some tests that are flaky and prone to failing

---

Checklist

• Does npm test pass with this change (including linting)?
• Is the new or changed code fully tested?
• Is a documentation update included (if this change modifies existing APIs, or introduces new ones)?

[bobymicroby]

Can you please add a tsdoc comment explaining the purpose behind isDirtyWatch / setDirtyWatch?

[bobymicroby]

Hi @htemelski 👋 We should use our docker container for spawning sentinel instances if we want to be able to switch/skip tests easily between server versions.

[bobymicroby]

Instead of adding dockerEnv, dockerBinds, and so on, please add them as optional properties to the RedisServerDockerConfig like :

export interface VolumeMount {
source: string; // Host path
target: string; // Container path
readonly?: boolean; // Read-only mount
}

export interface RedisServerDockerConfig {
image: string;
version: string;
env?: Record<string, string>;
volumes?: VolumeMount[];
}