Unlocking the Power of Application Monitoring

Introduction

• Today's applications are highly complex, involving multiple microservices, programming languages, frameworks, and components.
• This complexity presents a conundrum - developers and product teams want to ship products faster, while also ensuring quick detection and remediation of production issues.
• Application observability and monitoring can address this challenge by providing insights into the health, performance, and end-user experience of applications.

Why Application Observability?

• Understanding application health and gaining improved performance and availability for end customers.
• Reducing operational costs by having visibility into the application's inner workings.
• Improving end-user experience by proactively addressing issues before customers are impacted.

AWS Approach to Application Observability

• AWS offers a suite of products and services for application observability, including both open-source and AWS-native solutions.
• The AWS native approach starts with instrumentation using the Unified CloudWatch Agent, which collects metrics, logs, and traces with a single agent.
• These signals are then sent to CloudWatch, where you can create custom dashboards and enable cross-account observability.
• Additional insights are provided through AWS services like Container Insights, Lambda Insights, and Log Insights.

Application Signals

• Application Signals is AWS's opinionated way of performing Application Performance Monitoring (APM).
• It is based on the open-source OpenTelemetry standard, ensuring vendor-neutrality and avoiding vendor lock-in.
• Application Signals supports four programming languages: Java, Python, .NET, and Node.js, with automatic instrumentation through OpenTelemetry.
• It integrates with various compute platforms, including EC2, ECS, and EKS, as well as on-premise Kubernetes environments.

Key Features of Application Signals

1. Telemetry Collection:

   • Collects metrics, logs, and traces based on the OpenTelemetry standard, with backwards compatibility with X-Ray.
   • Leverages the Unified CloudWatch Agent for a single-agent approach.

2. Service Discovery:

   • Automatically maps the dependencies and relationships between your microservices.
   • Provides a dynamic service map, visualizing the correlations between services.

3. Troubleshooting:

   • Enables quick troubleshooting by providing a unified view of metrics, logs, and traces.
   • Allows users to navigate from anomalies in the graphs directly to the relevant transaction traces and logs.

4. Service Level Objectives (SLOs):

   • Supports the creation of SLOs based on service-level indicators (SLIs) and error budgets.
   • Simplifies the process of defining and monitoring SLOs, reducing the need for manual spreadsheet-based calculations.

5. Integrated Observability:

   • Integrates synthetic monitoring (Synthetics) and real-user monitoring (RUM) into the Application Signals dashboard.
   • Provides a holistic view of application performance, including both synthetic and real-user data.

Bridging the Gap between Applications and Infrastructure

• Application Signals maintains a tight integration with Container Insights, allowing users to seamlessly navigate between the application and infrastructure layers.
• For EKS environments, enabling the CloudWatch add-on is all that's required to set up Application Signals.

Demonstration

• The demo showcases the out-of-the-box dashboard provided by Application Signals, including:
  • Service status and red metrics (requests, errors, faults, and duration/latency)
  • Service map visualizing the dependencies between microservices
  • Detailed service-level information, including anomaly detection and troubleshooting capabilities
  • Service-level objective (SLO) creation and monitoring

Conclusion

• Application Signals is AWS's opinionated approach to Application Performance Monitoring (APM), leveraging the OpenTelemetry standard to provide vendor-neutral observability.
• It offers a comprehensive set of features, including telemetry collection, service discovery, troubleshooting, SLO management, and integrated synthetic and real-user monitoring.
• By bridging the gap between applications and infrastructure, Application Signals provides a holistic observability solution within the AWS ecosystem.

Additional Resources

• AWS Observability Workshop: Free hands-on workshop to explore best practices.
• Observability Accelerators: Terraform and CDK-based accelerators to jumpstart your observability journey.
• AWS Skill Builder Course: Free online course dedicated to AWS Observability.
• Additional Session: "Deep Dive on Application Signals" (COP409) on Thursday at 3:30 PM.