Feature_request: Timer Diagnostics for Active Queries

[Michealjunior]

Timer-driven behavior in this library is opaque: users cannot see which timers are active for their queries or when those timers will fire. Debugging unexpected refetches, stale transitions, or cache eviction requires guesswork. Provide an inspectable view of active timers and their next trigger time.

Design a read-only diagnostics API that returns a consistent snapshot of all active timers associated with queries and mutations. Include timers created for staleness, garbage collection, refetch intervals, and retries. Each diagnostic should describe timer kind, owning query, whether it repeats, and time until the next scheduled execution. Keep calls cheap, side-effect free, and safe for production use. Preserve existing behavior when diagnostics are never read.

Expose getTimerDiagnostics() on core client returning readonly diagnostics with fields: type (stale/gc/refetch/retry), queryHash, msUntilFire, isInterval.