Dynatrace is an enterprise observability platform built around a single agent, OneAgent, and an automatic dependency model called Smartscape. Install OneAgent on a host and it discovers every process, container, and service running there, instruments the supported runtimes without code changes, and maps how each service calls the others in real time. That auto-discovered topology is what sets Dynatrace apart from tools that expect you to define services and dashboards by hand - the platform builds the map for you and keeps it current as deployments change.

Davis and causal root-cause analysis

On top of the topology runs Davis, Dynatrace's analysis engine. Rather than firing a separate alert for every metric that crosses a threshold, Davis correlates anomalies across the dependency graph and tries to name a single root cause - a slow database is reported as one problem with the downstream services it affected listed as symptoms, not as a storm of fifty independent alerts. For a large estate with thousands of services, this collapsing of alert noise into ranked problems is the main reason teams choose it over assembling dashboards themselves.

Querying with DQL

Data lands in Grail, the platform's data store, and is queried with the Dynatrace Query Language (DQL), a pipeline syntax where each stage transforms the stream:

fetch logs
| filter dt.entity.service == "checkout"
| filter loglevel == "ERROR"
| summarize errorCount = count(), by:{bin(timestamp, 1m)}
| sort timestamp desc

Frequently asked questions

What is OneAgent? A single binary installed per host that auto-discovers and instruments processes, containers, and services, removing the need to deploy a separate agent per data type or per language.

What does Smartscape do? It builds and continuously updates a topology map of every entity and how they depend on each other, which Davis then uses to correlate problems.

How does Davis reduce alert noise? By analyzing anomalies against the dependency graph and reporting one ranked root-cause problem with its downstream symptoms attached, instead of an independent alert per metric.

Who is Dynatrace aimed at? Large enterprises with sprawling, fast-changing estates where manual dashboard and service definition does not scale and automatic discovery plus root-cause ranking pays off.

See the monitoring & logging overview and compare with AppDynamics.