FlagPilot

FlagPilot resolves flag evaluations locally inside your application processes. This architecture provides sub-millisecond evaluation times with zero database round-trips.

Evaluation Flow Architecture

When your code calls flagpilot.evaluate(), the SDK resolves the result using local state caching:

1. Read local Edge Config cache stored in memory.

↓

2. Validate environment toggle status: if disabled, resolve false.

↓

3. Target user verification: check if userId matches explicit targeting array.

↓

4. Hashing check: run Sha256 bucket allocation if percentage rollout is active.

↓

5. Return result and asynchronously log telemetry health metrics.

Caching & Configuration Sync

SDKs maintain a connection to Vercel's Edge Config (or run background polling triggers). Modifications made via the FlagPilot dashboard propagate globally in under 300 ms, ensuring your client instances are always in sync.

FAQ

How does Evaluation Flow interact with local caches?

Configurations are synced down using Edge Config or fallback REST triggers. The client automatically reads evaluations in-memory, resulting in sub-millisecond local evaluation times.

Where can I obtain support for Evaluation Flow?

You can open a support ticket in your dashboard, check the community roadmap, or consult the FlagPilot Platform Status page.