Feature Request: Allow overriding endpoint for HTTP resolver

Hello,

I'm working on an AppSync API which needs to talk to 3d party REST API.

The REST API URI has a following schema: `./api/
`
`` will depend on AppSync query/mutation input parameters.

Templates logic will be absolutely identitcal for all datasources.

So in current configuration, one would need either:

1) create one datasource per REST API, one query/mutation & template per one datasource and FE needs to be aware which query/mutation to call.

or

2) create one datasource per REST API, one query/mutation with bunch of pipeline resolver functions chained together with only one function being executed and the rest of them being skipped every time mutation/query is invoked.

Would be ideal if we can override `endpoint` similar to how it's currently done with DynamoDB batch get/put items templates where datasource table is ignored in favor of the table specified on the template. This was `endpoint` can be parametrized and decided during resolver execution.

Start by reviewing the HTTP resolver and endpoint handling, then compare them with the DynamoDB batch get/put templates referenced in the issue. Trace how datasources, resolver templates, and endpoint values are processed during resolver execution. Done means the endpoint can be selected from resolver input without duplicating datasources or templates, with documented coverage for the behavior.

Written by the indexing model from the issue text.