AWS DevOps Agent Builder — May 2026
Practical walkthrough of five agentic SRE challenges (onboarding, investigation, stress-testing, bad-deploy detection, and automation) with reproducible steps and outputs
AWS DevOps Agent Builder Challenge - May 2026
I completed the AWS DevOps Agent Builder Challenge for May 2026 and documented my work in this repository:
This challenge was a great hands-on way to explore DevOps, observability, incident investigation, automation, and agentic SRE workflows. The repository contains my solutions, findings, and notes for each challenge.
What I Built
The challenge is organized into five parts. Each part focuses on a different DevOps or SRE scenario, starting from basic agent setup and moving toward automated incident response.
Challenge 1 - Meet Your Agent
The first challenge introduced the agent environment and helped me understand how the agent works.
In this challenge, I verified the setup, explored the provided files, and checked how the agent reports its status. This helped me get familiar with the structure of the challenge and the expected workflow.
Key learning:
- Understanding the agent setup
- Reviewing the challenge structure
- Checking basic telemetry and logs
- Learning how the agent fits into a DevOps workflow
Challenge 2 - First Investigation
The second challenge focused on investigating an issue using logs and available signals.
I reviewed the failure scenario, checked the logs, and followed the evidence to identify the possible root cause. This challenge showed how important structured investigation is during an incident.
Key learning:
- Reading logs carefully
- Looking for error patterns
- Connecting symptoms with possible causes
- Writing clear findings after investigation
Challenge 3 - Stress and Diagnose
The third challenge involved stress testing and diagnosing system behavior under load.
I observed how the application behaved during increased traffic and reviewed the metrics to understand where performance started to degrade. This helped me practice identifying bottlenecks from system signals.
Key learning:
- Running a load scenario
- Watching latency and error behavior
- Identifying performance bottlenecks
- Connecting metrics with user-facing impact
Challenge 4 - Bad Deploy Detective
The fourth challenge focused on detecting a bad deployment.
I compared the system behavior before and after deployment changes and used the available evidence to find the risky change. This challenge was useful for understanding how deployments can affect reliability.
Key learning:
- Comparing behavior across deployments
- Looking for error rate changes
- Identifying suspicious changes
- Thinking about rollback and mitigation strategies
Challenge 5 - Build Your Own Agentic SRE
The final challenge was about building an agentic SRE workflow.
I worked on an automated response pattern where the agent observes a problem and suggests or performs a remediation step. This was the most interesting part because it connected monitoring, decision-making, and automation.
Key learning:
- Designing a simple incident response workflow
- Defining useful thresholds
- Thinking about safe automation
- Adding guardrails for production-style remediation
Main Takeaways
This challenge helped me understand how DevOps and SRE practices can be improved with agentic workflows. Instead of only reacting manually to incidents, an agent can help collect evidence, detect patterns, and support remediation.
My biggest takeaways were:
- Good observability makes investigation faster.
- Logs and metrics are most useful when they are connected to a clear incident timeline.
- Automation should be helpful, but it also needs safety checks.
- Agentic SRE can reduce response time when designed carefully.
Repository
All challenge solutions and notes are available here:
Conclusion
The AWS DevOps Agent Builder Challenge was a practical learning experience for DevOps automation and incident response. It helped me practice investigation, diagnosis, deployment analysis, and agent-based remediation in a structured way.
I enjoyed completing the challenge and building a better understanding of how agentic systems can support modern DevOps and SRE teams.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article