Skip to content

07-deployment-and-production-operations.md

Latest commit

 

History

History
63 lines (43 loc) · 2.87 KB

07-deployment-and-production-operations.md

File metadata and controls

63 lines (43 loc) · 2.87 KB
layout default
title Chapter 7: Deployment and Production Operations
nav_order 7
parent Strands Agents Tutorial

Chapter 7: Deployment and Production Operations

Welcome to Chapter 7: Deployment and Production Operations. In this part of Strands Agents Tutorial: Model-Driven Agent Systems with Native MCP Support, you will build an intuitive mental model first, then move into concrete implementation details and practical production tradeoffs.

This chapter outlines production rollout and operational governance for Strands agents.

Learning Goals

  • prepare Strands services for production workloads
  • build observability around agent and tool calls
  • handle incident and rollback scenarios
  • enforce security and dependency hygiene

Operational Checklist

  • pin versions for SDK, tools, and models
  • capture structured logs/metrics for tool behavior
  • define timeout/retry policies per integration
  • document runbooks for degraded dependencies

Source References

Summary

You now have a deployment and operations baseline for production Strands usage.

Next: Chapter 8: Contribution Workflow and Ecosystem Extensions

Source Code Walkthrough

Use the following upstream sources to verify deployment and production operations details while reading this chapter:

  • src/strands/telemetry/ — the observability module providing OpenTelemetry tracing and metrics export for production monitoring of agent invocations, tool calls, and model latency.
  • src/strands/telemetry/metrics.py — the metrics collection implementation that records token counts, latency, and tool call frequency as OTEL metrics for dashboarding and alerting.

Suggested trace strategy:

  • review src/strands/telemetry/tracer.py to see how OpenTelemetry spans are created around agent invocations and tool calls
  • trace metrics.py to understand which metrics are emitted by default and how to extend with custom metrics
  • check environment variable documentation for OTEL_EXPORTER_OTLP_ENDPOINT and related settings that control telemetry export in production

How These Components Connect

flowchart LR
    A[Agent invocation in production] --> B[telemetry/ OTEL tracing]
    B --> C[Spans sent to OTLP endpoint]
    C --> D[Dashboards and alerts]
    B --> E[metrics.py token and latency counts]
    E --> D
Loading