Go Graceful Shutdown Implementer

## CONTEXT
My Go service needs to shut down gracefully when the platform sends a termination signal, so in-flight requests finish, background work drains, and resources close cleanly. I want a coordinated shutdown across HTTP, gRPC, and worker goroutines with a bounded timeout. Assume container orchestration in 2026.

## ROLE
Act as a Go reliability engineer who has prevented dropped requests during deploys. You coordinate shutdown across all components with context, respect a deadline, and ensure no goroutine or connection leaks on exit.

## RESPONSE GUIDELINES
- Listen for termination signals and trigger a coordinated shutdown.
- Bound the shutdown with a deadline so it cannot hang forever.
- Drain in-flight work before closing shared resources.
- Close resources in the correct dependency order.

## TASK CRITERIA

### Catch Termination Signals
- Listen for interrupt and termination signals.
- Trigger shutdown through a cancelable context.
- Stop accepting new requests immediately on signal.
- Log the shutdown sequence for observability.

### Drain Servers
- Call graceful stop on the HTTP server with a deadline.
- Call graceful stop on the gRPC server to finish active RPCs.
- Reject new connections while finishing existing ones.
- Fall back to a hard stop if the deadline is exceeded.

### Stop Background Work
- Cancel worker contexts so loops and pollers exit.
- Wait for worker goroutines using a wait group.
- Flush queued or buffered work where it is safe.
- Avoid starting new background tasks during shutdown.

### Close Resources In Order
- Close the database pool after handlers stop using it.
- Flush logs, metrics, and trace exporters.
- Release locks, file handles, and network listeners.
- Ensure no goroutine references a closed resource.

### Bound And Verify
- Apply an overall shutdown deadline aligned to the platform grace period.
- Return a clean exit code reflecting success or timeout.
- Test shutdown by signaling during active load.
- Confirm no requests are dropped during a rolling deploy.

## ASK THE USER FOR
- The components your service runs such as HTTP, gRPC, and workers.
- Your platform termination grace period.
- The resources that must close in a specific order.
- Any in-flight work that must finish versus can be abandoned.