Documentation about default pool is somewhat confusing

[markmeeus]

Hi There,

We recently had some issues regarding the pooling functionality, and the documentation seemed a bit confusing.
From the code it looks like the behaviour is as follows:

When using {pool, default} tuple => the default pool will be used.
When not passing any pool tuple, the use_default_pool environment key is checked. If true, the default pool is used, if not, it does not use the pool.

So this statement from the readme

By default all connections are created and closed dynamically by hackney but sometimes you may want to reuse the same reference for your connections. It's especially useful if you just want to handle serially a couple of requests.

Seems a bit off, since the only way to dynamically create/close connections would be to not pass any pool option and switching off the default pool using the environment key.
By default it would use the default pool, which totally makes sense.

Also this:
Note:

Sometimes you want to disable the default pool in your app without having to set the client option each time. You can now do this by setting the hackney application environment key use_default_pool to false.

is a bit confusing because there is no way to switch off the default pool using a client option.

If my observations are correct, I would be happy to provide a PR for the README.

[sonic182]

I think this could be reopen, it is not clear the default configs for default pool, which is the max_connections for that pool?, timeouts? etc.