I’m in Austin working on a small internal tool that we’re building with an AI app builder rather than hand-coding, and I keep wondering how much observability we lose compared to a normal codebase. The tool will call a few internal APIs and we want to send traces and logs into Dynatrace so we’re not blind after launch.
What I’m unsure about is how much control these no-code builders give over the instrumentation side. Some let you add custom code or environment variables, but I don’t know if that’s enough to hook in the OneAgent or send OTLP data properly. There’s also the question of whether generated code is stable enough that the agents don’t get confused by frequent redeploys.
Has anyone here wired a no-code or AI-generated app into Dynatrace and gotten useful traces out of it? Curious what approach worked and where the gaps were.