Monitoring

getState(key) returns a non-mutating snapshot of a circuit, so you can poll it from a health endpoint or scheduled job without affecting behaviour (it never flips OPEN → HALF_OPEN).

import { Injectable, Inject } from '@nestjs/common';
import { CIRCUIT_BREAKER_SERVICE, ICircuitBreakerService, CircuitState } from '@nestjs-redisx/circuit-breaker';

/**
 * Expose circuit state for dashboards / health checks. `getState` is
 * non-mutating, so polling it never affects the breaker.
 */
@Injectable()
export class BreakerMonitor {
  private readonly circuits = ['stripe', 'users-api', 'search'];

  constructor(
    @Inject(CIRCUIT_BREAKER_SERVICE)
    private readonly breaker: ICircuitBreakerService,
  ) {}

  async snapshot(): Promise<Record<string, CircuitState>> {
    const entries = await Promise.all(this.circuits.map(async (key) => [key, (await this.breaker.getState(key)).state] as const));
    return Object.fromEntries(entries);
  }

  async isDegraded(): Promise<boolean> {
    const states = await this.snapshot();
    return Object.values(states).some((state) => state !== 'closed');
  }
}

Snapshot fields

interface ICircuitSnapshot {
  state: 'closed' | 'open' | 'half-open';
  failuresInWindow: number; // CLOSED: failures still inside the rolling window
  halfOpenSuccesses: number; // HALF_OPEN: probes that have succeeded
  halfOpenInFlight: number; // HALF_OPEN: permitted probes not yet resolved
}

Ideas

• Health check — report degraded when any circuit is not closed.
• Dashboards — expose per-circuit state as a gauge (0 = closed, 1 = half-open, 2 = open).
• Alerts — page when a critical circuit stays OPEN longer than expected.

TIP

Pair this with the Metrics plugin to publish circuit state as Prometheus gauges alongside your other application metrics.