[11.x] Add support for phpredis backoff and max retry config options

[TheLevti]

This pull request adds support for the phpredis backoff/retry configs.

Those settings are crucial for high performance applications. Currently I am forced to overwrite the driver in many services to leverage those configs. It would be amazing if you can merge this to support those configs out of the box.

I have confirmed with the phpredis maintainer, that those settings are currently only available to the normal Redis class, not for RedisCluster yet, so I did not add them to the cluster setup.

Reasoning

This has been added to phpredis by recommendation from an AWS blog post: https://aws.amazon.com/blogs/architecture/exponential-backoff-and-jitter/

Especially when using phpredis with AWS Serverless where auto scaling constantly drops connections, it is important to spread out the reconnect attempts. The default phpredis is a linear backoff approach, which is not ideal.

[Tklaversma]

Shouldn't we add the default values to the actual config file, so people can see those and adjust accordingly?

[TheLevti]

Shouldn't we add the default values to the actual config file, so people can see those and adjust accordingly?

Default is those values are not set at all. You can of course add the values that indicate disabled, but that is just unecessary work with no effect. Besides, not every setup has the Redis extension installed and that would lead to referencing unknown constants.

Those options are stated in the docs.