If you already monitor your web and mobile frontends with Real User Monitoring Classic, upgrading to Real User Monitoring touches every layer of your setup, from how RUM captures and processes your data to the measurements, analysis, and operations you build on top of it. Work through this upgrade guide with its sub-pages to complete the upgrade.
Real User Monitoring (RUM) gives you capabilities that RUM Classic cannot provide.
Experience Vitals,
Users & Sessions, and
Error Inspector provide task-oriented workflows for performance monitoring, user complaint resolution, and error investigation.
Dashboards,
Notebooks,
Workflows: Use your RUM data anywhere on the Dynatrace platform without exporting or bridging data.frontend.name or dt.security_context using Grail IAM policies. RUM Classic had no equivalent fine-grained control.RUM isn't a new interface on top of RUM Classic data. It captures, stores, and queries frontend data differently. That's why your dashboards, metrics, queries, and alerts don't carry over one-to-one.
The following sections summarize the most noteworthy new concepts, each with a short note on how it differs from RUM Classic and a link to the details. Knowing these concepts before you start helps you understand what changes in each upgrade step and why.
A frontend represents one end-user interface and carries its configuration, data grouping, and access scope. Classic called this an application. See RUM frontends.
Everything RUM captures, such as page loads, requests, errors, and user interactions, is stored as a user event in the user.events table. In Classic, the user action was the central record and everything else hung off it. See RUM data model.
Every event carries characteristics fields that describe what it contains, for example, a navigation, a request, an error, or a user interaction. One event can have several characteristics at once, so you find the events you need by filtering on these fields rather than on a single event type. See RUM data model.
A user action represents a significant frontend operation and its downstream effects. It starts with a trigger, such as a user interaction, a page load or navigation, an app start, or a call to the Dynatrace API, and covers the requests and other events that result from it. Those events stay separate but remain linked to their user action, so to see everything an action caused, you bring the action and its related events together in your analysis. See RUM data model.
Sessions land in the user.sessions table with no cap on the events they hold. Classic split a session after 200 user actions, so expect your RUM session counts to differ from your Classic counts. See RUM data model.
Custom key-value pairs enrich events and sessions with business context. These properties have their own configuration and don't inherit the Classic one. See RUM data model.
Web A page covers one document lifecycle, from one full page load to the next, and a view covers the content between two navigations. Each soft navigation in a single-page app starts a new view with its own summary, and the page summary covers all views of that document. Classic recognized only full page navigations. See Pages, views, and navigations.
Mobile Each visit to a screen is a view with its own summary, and transitions between views are navigation events. App starts are classified as cold, warm, or hot. Classic surfaced all three as autogenerated Loading … actions. See Concepts for mobile frontends.
RUM runs alongside Classic, which has to stay active, and your frontends send both beacon formats to the same endpoint. Most settings are shared, so a change on one side shows up on the other. See Upgrade data capturing.
Frontend-backend linking relies on the traceparent and tracestate headers. Classic used x-dtc and x-dynatrace, so you need new pass-through and CORS rules. See Frontend-backend linking.
You query user.events and user.sessions with DQL, and a single query can join frontend data with traces, logs, and metrics. USQL works only with Classic, so you need to rebuild every USQL query in DQL. See Dynatrace Query Language, Monitor web performance with DQL, and Analyze user behavior with DQL.
RUM ships built-in metrics under the dt.frontend prefix. Classic built-in metrics have functional replacements rather than one-to-one equivalents, see Built-in RUM metrics and Upgrade from classic RUM metrics to Grail. For custom metrics, you set up extraction from user events or user sessions in OpenPipeline.
A bucket sets the retention, storage tier, and query scope for the RUM data you route into it. Classic came with a fixed session retention you couldn't change. See Manage buckets for RUM.
Grail IAM policies grant read access record by record, based on values the records carry, such as which frontend the data came from. Classic had no comparable fine-grained control, and management zones were the closest equivalent. See RUM permissions and Segments.
Experience Vitals,
Users & Sessions, and
Error Inspector each cover one job: frontend health, session analysis, and error investigation. All three cover web and mobile frontends together, and you can take any analysis further in
Notebooks or
Dashboards.
Health alerts watch key frontend metrics and raise problems from anomalies, and you choose the anomaly detection analyzer per alert. Classic relied on application-level problem detection sensitivity. See Health alerts for frontends.
Dashboards,
Notebooks,
Workflows, Davis, and the Dynatrace MCP server read the RUM tables directly. Classic data needed an export or the REST API to be used anywhere else. See Analyze and alert with RUM.
Make sure that you meet the prerequisites, check which of your frontends are in scope, and assess how ready your RUM Classic setup is for the upgrade.
Before starting the upgrade, ensure the following are in place.
DPS license
RUM requires a Dynatrace Platform Subscription (DPS) license. If you are on a Classic license, contact your Dynatrace account team before proceeding.
When upgrading an existing frontend to RUM, Dynatrace protects you from double billing—Dynatrace charges you for ingested user sessions based on Dynatrace Classic and for the maximum number of properties configured in either Dynatrace Classic or Latest Dynatrace, not the sum of both.
Permissions
| Permission area | What you need |
|---|---|
| Settings | Write access to builtin:rum* schemas to configure frontends and instrumentation. Also builtin:health-experience.frontend-alert to manage health alerts. |
| Data access | storage:user.events:read, storage:user.sessions:read, storage:metrics:read, storage:smartscape:read |
| App access | app-engine:apps:run for Experience Vitals, Users & Sessions, and Error Inspector |
For a complete permissions reference, see RUM permissions.
You don't have to upgrade all your frontends at once. You can enable RUM on individual frontends and upgrade the rest at your own pace.
This upgrade guide covers the following frontend types:
| Frontend type | Coverage |
|---|---|
| Web frontends | Fully covered: browser-based monitoring via RUM JavaScript auto-injected by OneAgent or instrumented via manual insertion. |
| Mobile: Android, iOS, React Native, Flutter, .NET MAUI | Fully covered: native monitoring via the bundled RUM library for each platform. |
Currently unsupported and out-of-scope cases
The following are not covered by this upgrade guide. If your setup includes them, do not proceed with the upgrade for those frontends.
For all unsupported cases, contact your Dynatrace account team for roadmap guidance.
Before starting, use the Check your upgrade readiness dashboard to inventory your existing RUM Classic setup (frontends, instrumentation, settings, and agent versions) and identify any scope covered by the unsupported cases above.
To open the dashboard, go to
Dashboards > All dashboards > Ready-made and select Check your upgrade readiness.
Follow the steps below to upgrade your RUM setup.