[Software Arch Live Conf] Multi Cloud Observability using OTel

In this presentation we talk about:

As cloud adoption grows—with Gartner predicting 50% of enterprises will use the cloud by 2028 and 98% already considering multi-cloud strategies—organizations are moving from on-premises to hybrid and eventually to multi-cloud environments. The motivations include:

• Data Sovereignty: Complying with regional data laws.

• Vendor Lock-In Avoidance: Reducing dependency on any single cloud provider.

• Cost Optimization: Leveraging pricing advantages across clouds.

• Observability: Maintaining SLA/SLO visibility across distributed systems.

However, observability in multi-cloud setups becomes complex due to differing SDKs, schemas, and tools across vendors (e.g., AWS CloudWatch, Azure Monitor, Oracle Monitoring). This leads to operational overhead and fragmented on-call experiences.

Solution:
OpenTelemetry offers a vendor-agnostic, open-source standard for emitting metrics, logs, and traces. It supports major cloud providers and integrates seamlessly through exporters. Developers only need to use the OpenTelemetry SDK and configure exporters per provider.

Example:
In C#, adding new exporters (Azure, Prometheus, OTLP) becomes straightforward—promoting reuse and reducing complexity.

Visualization Tools like Grafana (for metrics) and Jaeger (for traces) unify data into a single pane of glass, further simplifying monitoring across cloud platforms.