A hands-on experiment attaching the OpenTelemetry Java agent to Apache NetBeans IDE instead of a production service. By patching netbeans.conf with three JVM flags and one config file change, the author routes JVM metrics, outbound HTTP spans, and bridged log records to Dash0. The result: real heap usage breakdowns by GC pool, thread counts, startup HTTP calls to update servers, deprecation warnings, and exit codes — all without writing a single line of plugin code. The post argues this approach can be applied fleet-wide to track IDE startup regressions, detect plugin-induced thread leaks, and correlate developer-side events with application-side telemetry in the same observability backend.
Table of contents
The question nobody thinks to askWhy this is more than a curiosityThe setupWhat arrives in Dash0What this enablesWhat comes nextThe broader point79 Impressions