Comprehensive AWS Observability: Architecting a Robust SRE Monitoring Stack and Advanced CloudWatch Techniques Beyond Basic Monitoring

SRE teams operating in complex environments require observability solutions that extend beyond standard dashboards to effectively manage and maintain cloud workloads at scale. In this session, we will explore how SREs can build a comprehensive AWS observability stack with AWS CloudWatch, Amazon Managed Service for Prometheus and Grafana, and AWS X-Ray to achieve complete visibility across an AWS infrastructure. We will emphasize how to build alerting systems that fight alert fatigue while improving incident response speed and implementing predictive monitoring and anomaly detection capabilities. Participants will gain detailed insights into advanced CloudWatch capabilities, including log insights, custom metrics, anomaly detection, application performance monitoring (application signals), cross-account observability, etc. Attendees will learn best practices for architecting observability explicitly tailored for SRE use cases. It will help attendees to learn how to capture key performance indicators (KPIs), automate responses and incident management workflows, and correlate metrics, logs, and tracing data to pinpoint issues quickly and reduce mean-time-to-resolution (MTTR).

Enhancing AWS CloudTrail Observability Using Event-Driven Ansible

AWS CloudTrail is a useful tool that records everything happening in AWS account, such as who did what, when, and from where. It helps with tracking changes, spotting unusual behavior, and meeting security and audit requirements. By combining CloudTrail with Event-Driven Ansible, we can take automatic action as soon as something important happens without needing someone to step in manually. In this talk, I will demonstrate how to set up this integration using the ansible.eda.aws_cloudtrail module. In this demo, I will give a demo on how to monitor events like starting or stopping an EC2 instance, and automatically trigger a response. This setup can be customized to handle many other events, helping teams manage their cloud systems more easily, securely, and efficiently.

Lambda Layers and Extensions: The unsung hero for observability

AWS Lambda Functions has been the core of Serverless compute and as a service as capabilities to extend additional reusable components with Lambda Layers and Lambda Extensions which enables not just one Lambda function but many more in your application to leverage additional dependencies and bringing in standardisation for observability and configurations. In this session, you will see how easy it is to use Lambda Layers and Lambda Extensions for observability with tools such as New Relic and even open-source Layer and Extensions by AWS Distro for OpenTelemetry (ADOT).

Crushing Bottlenecks: Distributed Load Testing on AWS

Building reliable applications in the cloud starts with understanding how they perform under pressure. This session dives into how AWS customers simulate real-world traffic and incorporate performance and load testing both pre- and post-deployment to validate system reliability under stress and varying load conditions. We'll cover testing APIs, microservices, and applications, and show how performance data drives smarter scaling, monitoring, and architectural decisions.

Intelligent Observability and Alerting with Amazon GenAI Services

As cloud workloads scale, traditional alerting often creates noise and delays incident response. This session explores how AWS Generative AI services bring intelligence to observability by enabling smarter alarm creation, proactive detection, and context-aware incident management. Learn how Amazon Q CLI, Amazon Bedrock, and the Strands SDK enhance CloudWatch alarms and integrate into incident response workflows. Walk away with practical patterns to reduce alert fatigue, accelerate resolution, and strengthen operational resilience.

What we learned (the hard way): Reconciliation at scale with serverless architecture

Scaling reconciliation in a financial system presents unique challenges, especially when processing massive volumes of transactional data. In this talk, we'll share how we designed and built a scalable, serverless reconciliation system using tools like DynamoDB, AWS Lambda, ECS Tasks and Step Functions. We'll walk through the architectural decisions, design patterns, and lessons we learned the hard way—covering topics such as data consistency, orchestration, error handling, and performance tuning. Attendees will leave with practical insights into building resilient, cost-effective financial workflows using a modern serverless stack.

Assemble a low cost FinOps Toolchain using AWS native solutions

In this talk I'll share how I moved our organization from an expensive third party tool to AWS native solutions for cost allocation and cost optimization. I'll share specifics on the tools we are using and some of the FinOps successes we've had.

What I learned using kiro during the preview phase

I was able to get on the preview of kiro before it was available to the public, and I provided a lot of feedback to the kiro developers before the GA. I will talk about my experiences using kiro as it was being built, and how it evolved into what it is today.

Harnessing the Power of AWS Network Firewall: Securing Your Egress Traffic with Centralized Control

Firewalls are crucial for infrastructure security as they help protect outgoing data from an organization's network, known as egress traffic. This session will cover the importance of securing egress traffic and explore how AWS Network Firewall can provide a centralized solution for managing and monitoring outbound traffic from your AWS environment.

Agentic AI on AWS: Building Agentic Systems for Enterprise Solutions

The emergence of agentic AI represents a transformative shift in how organizations leverage artificial intelligence for business operations. This session explores the implementation of AI agents on AWS, focusing on how these autonomous systems can enhance productivity and decision-making processes. We'll examine the current landscape of agent frameworks, from simple task automation to complex reasoning systems, and demonstrate how AWS's suite of services enables scalable agent deployment. The presentation will cover key aspects of the agent maturity curve, exploring the progression from basic rule-based agents to sophisticated autonomous systems capable of complex problem-solving. Attendees will learn about various implementation approaches, best practices for agent development on AWS infrastructure, and real-world use cases demonstrating successful enterprise integration. Special attention will be given to emerging frameworks, security considerations, and the strategic positioning of agentic AI solutions for different business scenarios.

Beyond Chat: Building Hybrid Agentic Workflows on AWS

While generative AI has revolutionized chat applications, its true potential lies in automating complex business processes. This session moves beyond conversational interfaces to explore the power of hybrid agentic workflows, which strategically blend deterministic logic with AI-driven decision-making. We will explore the spectrum of agentic integration: from a fully controlled, rule-based system to a semi-autonomous hybrid, and finally to a fully agentic workflow where the AI handles query generation, decision-making, and content creation. This approach enables choosing the correct agentic level for your application while keeping flexibility for changing it in future depending on your needs. We'll demonstrate how to build these sophisticated workflows using the new Strands Agents SDK from AWS and using services like AWS Lambda, Amazon Bedrock, and DynamoDB. You will leave with a clear understanding of how to identify opportunities for agentic automation in your own systems and a practical framework for designing, building, and deploying these next-generation applications on AWS.

Guardrails: Guardians of the AI Galaxy?

Guardrails are mentioned very often in AI conversations and even getting pulled into cybersecurity problems when there are better techniques which might result in weak solutions and other risks. This presentation covers common types of guardrails, where are they are frequently used and misused, when there are better techniques for cybersecurity solutions minimizing risk, better compliance and governance. And how this could ultimately lead to a solid groundwork for safer and responsible AI. The presentation centers around the use of AWS AI Stack, Amazon Bedrock and Amazon SageMaker with the help of several open source AI guardrails, integrated and layered for a complete AI security solution.