One Kubernetes workload, one Dynatrace service — a live walkthrough of what Service Detection v2 actually changes, run against your own Latest Dynatrace tenant in about thirty minutes.
SDv1 creates a separate Dynatrace service for every REST controller class and every Kafka listener on a single Kubernetes workload. A Spring Boot app with two controllers and one consumer becomes four services in the UI, each carrying its own health, baselines, and alerts. SDv2 collapses that to one UNIFIED service entity and pushes the per-controller and per-queue detail onto metric dimensions. Four things to monitor become one, and every slice you used to navigate to is now a query away.
A single DQL query tells the whole story: four dt.entity.service entities under SDv1, one UNIFIED entity under SDv2. A second companion workload shows how OTEL_SERVICE_NAME behaves — as a prefix on every SDv1 fragment, and as the outright name on SDv2.
Three metric families coalesced as the Transactions column. Databases and queues surface as dimensions on the caller, not as separate service entities. Metric-first DQL skips the entity-table join entirely. Plus a look at what changes when SDv2 for OneAgent reaches GA in June 2026.
| Component | What it is |
|---|---|
app/ | A small Spring Boot app — two REST controllers, a Kafka consumer, a JDBC client. |
k8s/ | Manifests for two namespaces on one GKE cluster, SDv1 and SDv2 side-by-side, with Postgres, Redpanda, and a k6 loadgen per side. |
presentation/ | Two-section narrative, Marp slides built to a single editable PPTX, and the unified sdv2-demo.yaml notebook with ten DQL questions. |
scripts/ | up.sh brings the stack up on GKE; load-demos.sh loads the notebook into your tenant with an environment share. |
# After you've set the DT_* env vars (see SETUP.md):
./scripts/up.sh # create the GKE cluster, deploy the app
./scripts/load-demos.sh # load the demo notebook into your tenantFull prerequisites — GKE, Dynatrace tokens, dtctl — are in the setup guide.
Each tag bundles the slide deck, the demo notebook YAML, and the full presentation tree on the releases page.