Monitor AWS Lambda

Why monitor AWS Lambda

Setting up AWS Lambda monitoring

1. 1
   Deploy the handler above as a dedicated /health function. Keep it separate from business logic so the health check survives main-app deploys.
2. 2
   Expose via Function URL or API Gateway. Function URLs are cheaper and skip the API Gateway 29-second timeout.
3. 3
   Configure provisioned concurrency (1 to 2) if cold starts matter for your use case. Otherwise expect sub-second cold starts for a simple handler, longer if initialization is heavy (large deps, Java/.NET, many security groups).
4. 4
   Add https://[url]/health in Hyperping at 30-second intervals. This also keeps the function warm for free.
5. 5
   Set alert threshold to 2 consecutive failures. Single failures are often throttling blips, not outages.
6. 6
   For background Lambda jobs (SQS, scheduled), use heartbeat monitors: the handler POSTs to a Hyperping heartbeat URL at the end of each successful run.

Example health endpoint

export const handler = async () => {
  const start = Date.now();

  try {
    // verify downstream dependencies
    await Promise.all([checkDynamo(), checkSecretsManager()]);
  } catch (err) {
    return {
      statusCode: 503,
      body: JSON.stringify({ status: 'degraded', error: err.message }),
    };
  }

  return {
    statusCode: 200,
    headers: { 'content-type': 'application/json' },
    body: JSON.stringify({
      status: 'ok',
      cold: !global.__warm,
      latency: Date.now() - start,
    }),
  };
};

global.__warm = true;

What typically goes wrong with AWS Lambda

Frequently asked questions