Fireflies.ai Observability

// lib/fireflies-instrumented.ts
import { Counter, Histogram, Gauge } from "prom-client";

const apiRequests = new Counter({
  name: "fireflies_api_requests_total",
  help: "Total Fireflies API requests",
  labelNames: ["operation", "status"],
});

const apiLatency = new Histogram({
  name: "fireflies_api_latency_seconds",
  help: "Fireflies API request latency",
  labelNames: ["operation"],
  buckets: [0.1, 0.25, 0.5, 1, 2, 5, 10],
});

const FIREFLIES_API = "https://api.fireflies.ai/graphql";

export async function firefliesQueryInstrumented(
  operation: string,
  query: string,
  variables?: any
) {
  const timer = apiLatency.startTimer({ operation });

  try {
    const res = await fetch(FIREFLIES_API, {
      method: "POST",
      headers: {
        "Content-Type": "application/json",
        Authorization: `Bearer ${process.env.FIREFLIES_API_KEY}`,
      },
      body: JSON.stringify({ query, variables }),
    });

    const json = await res.json();

    if (json.errors) {
      apiRequests.inc({ operation, status: json.errors[0].code || "error" });
      throw new Error(json.errors[0].message);
    }

    apiRequests.inc({ operation, status: "success" });
    return json.data;
  } catch (err) {
    apiRequests.inc({ operation, status: "failure" });
    throw err;
  } finally {
    timer();
  }
}

const webhookEvents = new Counter({
  name: "fireflies_webhook_events_total",
  help: "Webhook events received",
  labelNames: ["event_type", "status"],
});

const webhookProcessingTime = new Histogram({
  name: "fireflies_webhook_processing_seconds",
  help: "Time to process webhook events",
  buckets: [0.1, 0.5, 1, 5, 10, 30],
});

const transcriptQueue = new Gauge({
  name: "fireflies_transcript_queue_depth",
  help: "Number of transcripts queued for processing",
});

export async function handleWebhookWithMetrics(event: any) {
  const timer = webhookProcessingTime.startTimer();
  transcriptQueue.inc();

  try {
    await processTranscriptReady(event.meetingId);
    webhookEvents.inc({ event_type: event.eventType, status: "success" });
  } catch (err) {
    webhookEvents.inc({ event_type: event.eventType, status: "error" });
    throw err;
  } finally {
    timer();
    transcriptQueue.dec();
  }
}

const healthStatus = new Gauge({
  name: "fireflies_health_status",
  help: "Fireflies API health (1=healthy, 0=unhealthy)",
});

// Run every 5 minutes
async function healthProbe() {
  try {
    const start = Date.now();
    const data = await firefliesQueryInstrumented("health_check", "{ user { email } }");
    const latencyMs = Date.now() - start;

    healthStatus.set(1);
    console.log(`Fireflies health: OK (${latencyMs}ms)`);
  } catch (err) {
    healthStatus.set(0);
    console.error(`Fireflies health: FAILED - ${(err as Error).message}`);
  }
}

setInterval(healthProbe, 5 * 60 * 1000);

const seatUtilization = new Gauge({
  name: "fireflies_seat_utilization",
  help: "Transcripts per user",
  labelNames: ["user_email"],
});

const totalSeats = new Gauge({
  name: "fireflies_total_seats",
  help: "Total Fireflies seats",
});

// Run daily
async function trackSeatUtilization() {
  const data = await firefliesQueryInstrumented("seat_audit", `{
    users { email num_transcripts }
  }`);

  totalSeats.set(data.users.length);
  for (const user of data.users) {
    seatUtilization.set({ user_email: user.email }, user.num_transcripts);
  }

  const inactive = data.users.filter((u: any) => u.num_transcripts < 2);
  if (inactive.length > 3) {
    console.warn(`${inactive.length} seats with <2 transcripts -- review for cost savings`);
  }
}

# prometheus/rules/fireflies.yml
groups:
  - name: fireflies
    rules:
      - alert: FirefliesAPIDown
        expr: fireflies_health_status == 0
        for: 10m
        labels:
          severity: critical
        annotations:
          summary: "Fireflies API unreachable for 10+ minutes"

      - alert: FirefliesHighErrorRate
        expr: rate(fireflies_api_requests_total{status!="success"}[5m]) > 0.1
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Fireflies API error rate >10% over 5 minutes"

      - alert: FirefliesRateLimited
        expr: rate(fireflies_api_requests_total{status="too_many_requests"}[5m]) > 0
        labels:
          severity: warning
        annotations:
          summary: "Fireflies API rate limiting detected"

      - alert: FirefliesWebhookBacklog
        expr: fireflies_transcript_queue_depth > 50
        for: 15m
        labels:
          severity: warning
        annotations:
          summary: "Webhook processing backlog exceeds 50 transcripts"

      - alert: FirefliesSlowProcessing
        expr: histogram_quantile(0.95, rate(fireflies_webhook_processing_seconds_bucket[1h])) > 30
        labels:
          severity: warning
        annotations:
          summary: "Webhook processing P95 exceeds 30 seconds"