Introduction to Dynatrace
Understand Dynatrace's architecture, its auto-discovery engine, the OneAgent approach, and how Davis AI differentiates it from every other monitoring platform.
Simple Explanation (ELI5)
Dynatrace is a single platform that watches your entire stack automatically. You install one small agent on each host — it discovers every process, service, container, and database dependency without any configuration. It then maps all relationships and uses AI to identify root causes before your on-call team even gets paged.
What Makes Dynatrace Different
Most monitoring tools require you to manually configure dashboards and alert thresholds. Dynatrace does this automatically through three core differentiators: OneAgent (zero-config instrumentation), Smartscape (auto-discovered topology), and Davis AI (automatic root cause analysis).
Dynatrace Architecture
/ Cloud
(per host)
(proxy)
SaaS / Managed
+ UI
Core Components
A single deployable agent installed on each host. Automatically instruments processes, collects JVM/CLR bytecode telemetry, container metadata, network, and OS metrics. Zero configuration required.
Secure proxy between OneAgents and the Dynatrace cluster. Used for cloud API polling (AWS, Azure, GCP), extension framework, and running Synthetic monitors from private locations.
SaaS or self-managed backend that stores all telemetry, runs Davis AI analytics, and serves the Dynatrace UI and REST APIs.
Automatically updated topology model: hosts → processes → services → applications → users. Every entity and every dependency mapped in real time.
Causal AI engine. Baselines every metric automatically, detects anomalies, correlates related signals across the full stack, and produces a single root-cause problem card.
REST APIs for querying metrics/logs/events, configuration-as-code (Monaco), custom event ingestion, and CI/CD pipeline integration.
Full-Stack Monitoring Coverage
| Layer | What Dynatrace Monitors | How |
|---|---|---|
| Application | Services, transactions, code hotspots, errors, SLOs | OneAgent bytecode instrumentation |
| Infrastructure | CPU, memory, disk, network per host/container | OneAgent OS integration |
| Kubernetes | Cluster, node, pod, namespace health; workload metrics | OneAgent + K8s API extension |
| Cloud | AWS/Azure/GCP native services (RDS, Lambda, Blob) | ActiveGate cloud integrations |
| Real User Monitoring | Browser/mobile sessions, Apdex, JS errors, Core Web Vitals | Dynatrace JavaScript RUM agent |
| Synthetic | Scheduled HTTP/browser tests from global locations | Dynatrace Synthetics |
Deployment — OneAgent on Linux
# Download OneAgent installer from your Dynatrace environment wget -O Dynatrace-OneAgent.sh \ "https://your-env.live.dynatrace.com/api/v1/deployment/installer/agent/unix/default/latest\ ?Api-Token=YOUR_PAAS_TOKEN" chmod +x Dynatrace-OneAgent.sh sudo ./Dynatrace-OneAgent.sh \ --set-server=https://your-env.live.dynatrace.com \ --set-tenant=YOUR_ENV_ID \ --set-tenant-token=YOUR_TENANT_TOKEN # Verify agent is running sudo systemctl status oneagent
Dynatrace on Kubernetes — DynaKube Operator
apiVersion: dynatrace.com/v1beta1
kind: DynaKube
metadata:
name: dynakube
namespace: dynatrace
spec:
apiUrl: https://your-env.live.dynatrace.com/api
tokens: dt-tokens # Secret with apiToken and dataIngestToken
oneAgent:
cloudNativeFullStack: # Injects OneAgent into every pod automatically
tolerations:
- effect: NoSchedule
key: node-role.kubernetes.io/master
activeGate:
capabilities:
- routing
- kubernetes-monitoringAuto-discovery in Practice
Once OneAgent is installed, Dynatrace automatically discovers:
- All running processes and their technology stack (Java, .NET, Node.js, PHP, Go, etc.)
- All network connections between processes — mapped as service-to-service dependencies
- All databases, queues, and caches the application calls
- Kubernetes workloads, namespaces, and container metadata
- Cloud load balancers and managed services (if ActiveGate cloud integration is configured)
Dynatrace vs Competitors
| Feature | Dynatrace | Datadog | New Relic | Prometheus+Grafana |
|---|---|---|---|---|
| Auto-discovery | Full auto (OneAgent) | Agent + manual config | Agent + manual config | Manual scrape targets |
| AI root cause | Davis AI (causal) | Watchdog (correlation) | Applied Intelligence | None (manual) |
| Distributed tracing | PurePath (automatic) | APM (manual SDK) | Distributed Tracing | Jaeger / Tempo |
| Cost model | DPS (usage-based) | Per host + per metric | Data ingest GB | Open source |
| K8s support | Operator + full pod injection | Helm chart | Helm chart | kube-prometheus-stack |
Debugging Scenarios
- OneAgent installed but no data in UI: Check agent status with
systemctl status oneagent. Verify network path to Dynatrace cluster (outbound HTTPS/443). Check firewall rules blocking the agent. - Service not appearing on Smartscape: The process may not be on a supported technology list, or it may be running under a user that OneAgent doesn't have permission to monitor.
- Kubernetes pods not injected: DynaKube custom resource may be missing the namespace annotation or the Operator is not running. Check
kubectl get dynakube -n dynatrace.
Real-world Use Case
A retail company migrating from on-premises to cloud needed observability across 400+ microservices. Rather than configuring individual Prometheus exporters per service, they deployed Dynatrace OneAgent on every Kubernetes node via the Operator. Within 30 minutes, Smartscape had automatically mapped all 400+ services, their database dependencies, and cross-service call paths — a mapping exercise that had previously taken their platform team three weeks to do manually.
Interview Questions
Beginner
A single agent installed per host that automatically instruments all processes, collects code-level metrics and traces, and sends telemetry to the Dynatrace cluster — no per-service configuration needed.
Dynatrace's automatically maintained real-time topology map — visualizing every entity (host, process, service) and every dependency relationship discovered by OneAgent.
Dynatrace's causal AI engine that automatically baselines metrics, detects anomalies, correlates related events across the full stack, and pinpoints the root cause in a single problem card.
A Dynatrace component that acts as a secure proxy for OneAgent communication and enables cloud integrations (AWS, Azure, GCP API polling) and extension framework capabilities.
OneAgent communicates via HTTPS on port 443 to the Dynatrace cluster or ActiveGate — no inbound firewall rules required.
Intermediate
SaaS: Dynatrace hosts and manages the cluster. Managed: Customer hosts the Dynatrace cluster on-premises or in their own cloud — used when data sovereignty requirements prevent data leaving the organisation.
Via the Dynatrace Operator and DynaKube custom resource — OneAgent is injected as a DaemonSet or via admission webhook into pods. The Operator also calls the Kubernetes API to collect namespace and workload metadata.
Dynatrace's unified consumption-based licensing model — you purchase a pool of Davis Data Units and spend them across any capability (infrastructure, APM, logs, RUM) based on actual usage.
Monitoring-as-Code CLI tool that allows Dynatrace configuration (dashboards, alerts, SLOs, synthetic tests) to be stored in version control and deployed via CI/CD pipelines.
Infrastructure metrics, full-stack APM (service monitoring, PurePath traces), process-level visibility, log collection, real user monitoring injection, and network monitoring — all from a single installation.
Scenario-based
1. Verify OneAgent is running: systemctl status oneagent. 2. Check Dynatrace UI Host → Processes for the host. 3. Confirm the JVM is on a supported JDK version. 4. Ensure OneAgent has OS-level access to the JVM process. 5. Check OneAgent logs at /var/log/dynatrace/oneagent/.
Install the Dynatrace Operator via Helm, create a DynaKube custom resource with cloudNativeFullStack mode — OneAgent is automatically injected into existing and new pods, and Kubernetes metadata is collected within minutes.
Yes — either use Dynatrace SaaS with an EU region environment, or deploy Dynatrace Managed on-premises/in your own EU cloud account, where you fully control data storage and residency.
Use Monaco (Monitoring-as-Code) to define all Dynatrace settings in YAML/JSON, store in Git, and deploy through CI/CD pipelines with environment-specific variable substitution.
Auto-discovery with OneAgent requires zero per-service setup and works immediately for any supported technology. Prometheus requires manually adding exporters, configuring scrape targets, and writing recording rules per service — high ongoing maintenance at scale.
Summary
Dynatrace's value proposition is radical simplicity at scale: one agent, zero configuration, full-stack visibility. OneAgent auto-discovers your entire environment, Smartscape maps all dependencies, and Davis AI handles the analysis — letting your team focus on fixing problems rather than building and maintaining monitoring infrastructure.