Cloud-based Observability in Distributed Systems : Designing a scalable observability architecture in the cloud using Azure, OpenTelemetry and .NET

Sammanfattning

This thesis investigates the design and implementation of cloud-native observability in distributed .NET systems using OpenTelemetry and Microsoft Azure. The work addresses the challenges of instrumenting modern applications to collect, process and visualize telemetry signals - logs, metrics and traces. A Proof-of-Concept (PoC) observability architecture was developed with Azure Monitor, Application Insights and Managed Grafana to demonstrate best practices in telemetry pipeline design, Infrastructure-as-Code deployment and alert-based automation. The system includes custom metrics, trace correlation and simulated errors to validate end-to-end observability. The results demonstrate that OpenTelemetry, combined with Azure Monitor and Application Insights, enables reliable and scalable telemetry collection with real-time traceability, metric visualization and alert-driven automation in a .NET cloud environment. Additionally, ethical and environmental aspects of observability were discussed, including compliance with privacy laws and reducing telemetry overhead. The solution is in accordance with Microsoft’s observability models and contributes practical insights for enterprise systems aiming for scalable and maintainable observability strategies.