Managing microservice architectures requires navigating highly complex systems. This complexity informs how we approach making these systems observable.

We’ll review a standardized platform-focused approach to building effective observable architectures, including how it addresses the new organizational challenges specific to microservices. The platform-focused approach encompasses three parts: the patterns we use, the needed organizational support, and the stack we use.

- Define observability: “The ability to understand what’s happening inside of your software systems to debug problems you've never seen before just using the telemetry (traces, logs, & metrics) emitted by your applications. It's also not just about a specific tool, it's about a team's ability to analyze that telemetry data.” (Liz Fong-Jones, Honeycomb)
- Three new organization challenges of microservices & how they apply to observability: silos & drift, multiple points of failure, & inherent lack of certainty.
- Introduce the idea of a standardized observability platform that encompasses patterns, support, and a stack.
- Patterns: focused on standardized communication protocols with three patterns.
- Support Strategy: Tie back to the part of our observability definition “about a team's ability to analyze that telemetry data.”
- Observability Stack: Review the different parts of the stack, referencing ETL patterns from my data engineering background (telemetry collection, processing, and exporting, which makes up our “communication layer”) to enable utilizing observability tools like DataDog, Honeycomb, SumoLogic, etc.
- Communication Layer Tool Principles: Data reliability & richness, followed by best practices.
- Investigation & Analysis Tools: Since most organizations aren’t building their own investigation tools, I focus on effective decision criteria for selecting tools.