Building, Deploying and Observing SDKs as a Service
Manually maintaining SDKs across multiple programming languages is slow, error-prone, and a constant drag on developer velocity. This episode treats SDK generation as a service—automated, containerised, and fully observable.
What You’ll Learn
- Why SDKs reduce developer friction compared to direct REST API integration (handling auth flows, retry logic, and serialisation plumbing automatically)
- Using OpenAPI Generator (open-source community project) to automatically generate SDKs in 50+ languages from an OpenAPI/Swagger specification
- Customising SDK output using Mustache templates to match your organisation’s coding conventions and style
- Deploying the OpenAPI Generator as a containerised service for automated, pipeline-driven SDK generation triggered by API spec changes
- Testing SDKs at scale: smoke testing generated clients against a live API sandbox before publishing, and why Python is a useful early canary language
- Observability with OpenTelemetry: auto-tracing SDK invocations end-to-end without modifying application code (agentless instrumentation)
- Using Lumigo for distributed tracing and visualising the full invocation chain across microservices and serverless functions
- The importance of contributing back to open-source communities whose tools power your production systems
Guest
This episode features Steve Kuching, developer advocate and OpenTelemetry practitioner, discussing the intersection of SDK automation, observability, and the open-source ecosystem.
Related Content
Chaos Engineering for Cloud native Apps
Chaos engineering is the discipline of proactively experimenting on distributed systems to build confidence in their ability to withstand production failures. Chris is joined by Ashish Balgath (Cloud Solution Architect at Thoughtworks) to explore how to introduce fault-injection experiments incrementally — from a developer's local machine all the way to production — using tools such as Chaos Monkey and fault-injection simulator tools.
Tales from the Real World - Shift Left your Performance Tests
Performance testing has traditionally lived in high-fidelity pre-production environments — but by the time you find a performance issue that late in the cycle, the cost to fix it is already compounding. In this episode, Chris is joined by HariKrishnan, Cloud Transformation Consultant, to explore a practical shift-left approach to performance testing. Hari shares real-world experience identifying the categories of performance issues that can be reproduced locally or in CI pipelines using tools like Gatling and Docker, reducing MTTR and avoiding the environment time-sharing bottlenecks that slow teams down.
DevOps Trends
A decade after Patrick Debois coined "DevOps," the landscape looks radically different. In this episode, Daniela Fontani — CTO at Central Consulting and long-time open source contributor — breaks down the most important DevOps trends reshaping the industry today: DevSecOps, GitOps, NoOps, automation-first pipelines, and the growing role of platform engineering. Plus, the honest truth about which buzzwords actually matter and which you can safely ignore.