Serverless Tip #10: Don't forget about observability

When something in a distributed architecture inevitably goes wrong, you need to be able to identify the root cause quickly.

Having great observability means that you can understand an application’s internal state based on external outputs. These outputs are often grouped into three pillars:

• Structured logs
• Metrics
• Traces

The teams behind AWS Lambda Powertools have created libraries that make it easy to instrument your Lambda functions with structured logging, custom metrics, and tracing capabilities.

Powertools is available in many languages, such as Python, TypeScript, .NET, and Java.

And, if the three main utilities for logs, traces, and metrics aren’t enough, Powertools supplies many utilities for everyday use cases, such as parsing events, feature flagging, and idempotency.