[FLI-1123] Support bulk flag evaluation in OFREP

[markphelps]

Problem

Re: #2980

We have single flag evaluation in #3267 (thanks @thepabloaguilar !)

We should support bulk flag evaluation https://github.com/open-feature/protocol/blob/218fd82f1d37844f26ec22418655a44395fbd3f4/service/openapi.yaml#L84

Then we will be compatible with the current ofrep specification

Ideal Solution

support bulk flag evaluation https://github.com/open-feature/protocol/blob/218fd82f1d37844f26ec22418655a44395fbd3f4/service/openapi.yaml#L84 in Flipt OFREP server

Search

• I searched for other open and closed issues before opening this

Additional Context

No response

_FLI-1123

[markphelps]

One thing to think about re implementing this:

Flipt requires flagKey for each bulk evaluation request, whereas OFREP spec does not specify it. However, they do state that we can 'filter' which flags to evaluate in the request context: https://github.com/open-feature/protocol/blob/218fd82f1d37844f26ec22418655a44395fbd3f4/service/adrs/0004-no-flag-filter-in-bulk-evaluation-request.md

I could see us adding a flagKeys field in the context which would be a list of the flags we want to evaluate, then in the bridge we could copy the context for each request when mapping to the flipt.EvaluationRequest

So it would be a 1 to N mapping, one batch eval request on the OFREP side would map to N requests in the bridge, where N is the length of flagKeys array.

We could return an error INVALID_CONTEXT if no flagKeys field is present.

Similarly for the namespaceKey (provided in X-Flipt-Namespace header, we would copy it to each evaluation request on the Flipt bridge side.