[DOC] What problem does flagd solve?

[agardnerIT]

Change in the documentation

This issue to to track the "what problem does flagd solve?" bulletpoint here: https://github.com/orgs/open-feature/projects/5/views/1?pane=issue&itemId=22809077

[agardnerIT]

My Take

Users looking to adopt OpenFeature are fundamentally looking for a way to adopt feature flagging in a vendor neutral and open way. Users deciding to adopt the standard will look for a pre-made tool that conforms to "best practices" and demonstrates "what it looks like for it do be done correctly". Of course, this will not be an academic exercise, users need a tool that works too!

In other words, the first question would be: "why OpenFeature" and once they've decided to try OpenFeature, the next question will be "how OpenFeature". flagd is the easiest and also the open source answer to the second question.

UEFI has EDK 2. OpenTelemetry has Jaeger. OpenFeature has flagd.

flagd demonstrates how to perform flag retrieval (from a backend store), evaluation (using things like targeted evaluation and fractional evaluation) and making those available to the client-side OpenFeature SDK code.

Of course, much like Jaeger and Prometheus, flagd is not the only OpenFeature compliant "flag evaluation" engine. Users can (and probably will at some point) opt for a vendor who supports OpenFeature. But flagd is the way the world learns OpenFeature. It's how they get started. It's what they can use, without restriction or limitation - until they outgrow the solution - much like when companies move to vendor Observability tools over Jaeger and Prometheus.