Agent skill

linear-observability

Implement monitoring, logging, and alerting for Linear integrations. Use when setting up metrics collection, dashboards, or configuring alerts for Linear API usage. Trigger: "linear monitoring", "linear observability", "linear metrics", "linear logging", "monitor linear", "linear Prometheus", "linear Grafana".

Stars 1,803
Forks 241

Install this agent skill to your Project

npx add-skill https://github.com/jeremylongshore/claude-code-plugins-plus-skills/tree/main/plugins/saas-packs/linear-pack/skills/linear-observability

SKILL.md

Linear Observability

Overview

Production monitoring for Linear integrations using Prometheus metrics, structured logging with pino, health checks, and alerting rules. Track API latency, error rates, rate limit headroom, and webhook throughput.

Prerequisites

  • Linear integration deployed
  • Prometheus or Datadog for metrics
  • Structured logging (pino, winston)
  • Alerting system (PagerDuty, OpsGenie, Slack)

Instructions

Step 1: Define Metrics

typescript
// src/metrics/linear-metrics.ts
import { Counter, Histogram, Gauge, register } from "prom-client";

export const metrics = {
  // API request tracking
  apiRequests: new Counter({
    name: "linear_api_requests_total",
    help: "Total Linear API requests",
    labelNames: ["operation", "status"],
  }),

  // Request duration
  apiLatency: new Histogram({
    name: "linear_api_request_duration_seconds",
    help: "Linear API request duration",
    labelNames: ["operation"],
    buckets: [0.1, 0.25, 0.5, 1, 2, 5, 10],
  }),

  // Rate limit headroom
  rateLimitRemaining: new Gauge({
    name: "linear_rate_limit_remaining",
    help: "Remaining rate limit budget",
    labelNames: ["type"], // "requests" or "complexity"
  }),

  // Webhook tracking
  webhooksReceived: new Counter({
    name: "linear_webhooks_received_total",
    help: "Total webhooks received",
    labelNames: ["type", "action"],
  }),

  webhookProcessingDuration: new Histogram({
    name: "linear_webhook_processing_seconds",
    help: "Webhook processing duration",
    labelNames: ["type"],
    buckets: [0.01, 0.05, 0.1, 0.5, 1, 5],
  }),

  // Cache effectiveness
  cacheHits: new Counter({
    name: "linear_cache_hits_total",
    help: "Cache hit count",
    labelNames: ["key"],
  }),
  cacheMisses: new Counter({
    name: "linear_cache_misses_total",
    help: "Cache miss count",
    labelNames: ["key"],
  }),
};

// Expose metrics endpoint
app.get("/metrics", async (req, res) => {
  res.set("Content-Type", register.contentType);
  res.end(await register.metrics());
});

Step 2: Instrumented Client Wrapper

typescript
import { LinearClient } from "@linear/sdk";

function instrumentedCall<T>(
  operation: string,
  fn: () => Promise<T>
): Promise<T> {
  const timer = metrics.apiLatency.startTimer({ operation });

  return fn()
    .then((result) => {
      metrics.apiRequests.inc({ operation, status: "success" });
      timer();
      return result;
    })
    .catch((error: any) => {
      const status = error.status === 429 ? "rate_limited" : "error";
      metrics.apiRequests.inc({ operation, status });
      timer();
      throw error;
    });
}

// Usage
const client = new LinearClient({ apiKey: process.env.LINEAR_API_KEY! });

const teams = await instrumentedCall("teams", () => client.teams());
const issues = await instrumentedCall("issues", () =>
  client.issues({ first: 50 })
);

Step 3: Structured Logging

typescript
import pino from "pino";

const logger = pino({
  level: process.env.LOG_LEVEL ?? "info",
  formatters: {
    level: (label) => ({ level: label }),
  },
});

const linearLog = logger.child({ component: "linear" });

// Log API calls
function logApiCall(operation: string, durationMs: number, success: boolean, meta?: any) {
  linearLog.info({
    event: "api_call",
    operation,
    durationMs,
    success,
    ...meta,
  });
}

// Log webhook events
function logWebhook(type: string, action: string, deliveryId: string, meta?: any) {
  linearLog.info({
    event: "webhook",
    type,
    action,
    deliveryId,
    ...meta,
  });
}

// Log errors with context
function logError(operation: string, error: any) {
  linearLog.error({
    event: "error",
    operation,
    errorMessage: error.message,
    errorStatus: error.status,
    errorType: error.type,
    // Never log API keys or tokens
  });
}

Step 4: Health Check Endpoint

typescript
interface HealthCheck {
  status: "healthy" | "degraded" | "unhealthy";
  checks: Record<string, {
    status: string;
    latencyMs?: number;
    error?: string;
  }>;
  timestamp: string;
}

async function checkLinearHealth(client: LinearClient): Promise<HealthCheck> {
  const checks: HealthCheck["checks"] = {};

  // Check API connectivity
  const apiStart = Date.now();
  try {
    const viewer = await client.viewer;
    checks.linear_api = {
      status: "healthy",
      latencyMs: Date.now() - apiStart,
    };
  } catch (error: any) {
    checks.linear_api = {
      status: "unhealthy",
      latencyMs: Date.now() - apiStart,
      error: error.message,
    };
  }

  // Check rate limit headroom
  try {
    const resp = await fetch("https://api.linear.app/graphql", {
      method: "POST",
      headers: {
        Authorization: process.env.LINEAR_API_KEY!,
        "Content-Type": "application/json",
      },
      body: JSON.stringify({ query: "{ viewer { id } }" }),
    });
    const remaining = parseInt(resp.headers.get("x-ratelimit-requests-remaining") ?? "5000");
    metrics.rateLimitRemaining.set({ type: "requests" }, remaining);

    checks.rate_limit = {
      status: remaining > 100 ? "healthy" : "degraded",
      latencyMs: remaining,
    };
  } catch {
    checks.rate_limit = { status: "unknown" };
  }

  const overall = Object.values(checks).some(c => c.status === "unhealthy")
    ? "unhealthy"
    : Object.values(checks).some(c => c.status === "degraded")
    ? "degraded"
    : "healthy";

  return { status: overall, checks, timestamp: new Date().toISOString() };
}

app.get("/health/linear", async (req, res) => {
  const health = await checkLinearHealth(client);
  res.status(health.status === "unhealthy" ? 503 : 200).json(health);
});

Step 5: Alerting Rules (Prometheus)

yaml
# prometheus/linear-alerts.yml
groups:
  - name: linear
    rules:
      - alert: LinearHighErrorRate
        expr: |
          rate(linear_api_requests_total{status="error"}[5m])
          / rate(linear_api_requests_total[5m]) > 0.05
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Linear API error rate > 5%"

      - alert: LinearRateLimitLow
        expr: linear_rate_limit_remaining{type="requests"} < 100
        for: 2m
        labels:
          severity: critical
        annotations:
          summary: "Linear rate limit remaining < 100 requests"

      - alert: LinearHighLatency
        expr: |
          histogram_quantile(0.95, rate(linear_api_request_duration_seconds_bucket[5m])) > 2
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Linear API p95 latency > 2 seconds"

      - alert: LinearWebhookProcessingSlow
        expr: |
          histogram_quantile(0.95, rate(linear_webhook_processing_seconds_bucket[5m])) > 5
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Webhook processing p95 > 5 seconds"

Step 6: Webhook Instrumentation

typescript
// Instrument webhook handler
app.post("/webhooks/linear", express.raw({ type: "*/*" }), async (req, res) => {
  const start = Date.now();
  // ... signature verification ...

  const event = JSON.parse(req.body.toString());
  const delivery = req.headers["linear-delivery"] as string;

  metrics.webhooksReceived.inc({ type: event.type, action: event.action });
  logWebhook(event.type, event.action, delivery);

  res.json({ ok: true });

  try {
    await processEvent(event);
    metrics.webhookProcessingDuration.observe(
      { type: event.type },
      (Date.now() - start) / 1000
    );
  } catch (error: any) {
    logError("webhook_processing", error);
  }
});

Error Handling

Error Cause Solution
Metrics not collecting Missing instrumentation Wrap all client calls with instrumentedCall()
Alerts not firing Thresholds too high Adjust based on actual traffic patterns
Health check timeout Linear API slow Add 10s timeout to health check
Log volume too high Debug level in production Set LOG_LEVEL=info in prod

Examples

Quick Health Check

bash
curl -s http://localhost:3000/health/linear | jq .
# { "status": "healthy", "checks": { "linear_api": { "status": "healthy", "latencyMs": 150 } } }

Resources

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results