AI agents removed the friction from writing telemetry
I used to avoid adding telemetry because it felt like tedious busywork. Now with Claude handling the OpenTelemetry boilerplate, I'm instrumenting everything.
Semantic versioning is a way for library authors to communicate to consumers that you’ve released a new version and information about the kind of version you’ve published. I’ve written about this in more depth before, but here’s the gist:
When you're installing open-source packages, check the version number. The authors of the component library are trying to communicate details about the package's stability to you.
I used to avoid adding telemetry because it felt like tedious busywork. Now with Claude handling the OpenTelemetry boilerplate, I'm instrumenting everything.
AI agents work better when given appropriate context and guardrails.
I've been using Claude Code to offload tedious parts of platform engineering: dependency reviews, generating test PRs, dependency migrations, and project cleanup.