Dynatrace Release Radar 10.26

This series covers recent Dynatrace releases and updates, focusing on what's new, what's changed, and how these recent enhancements can benefit you and your organization. Each post covers newly available capabilities and where to explore them.

October's updates help you tackle the work that comes before a fix. The SRE Agent gathers evidence and reports its findings. Distributed Tracing keeps your filters and follows linked requests across queues. Experience scores and network views show you who's affected and which connections need a closer look.

Explore October's updates on our Release Radar launchpad on the Dynatrace Playground.

Give agentic AI the first pass at a problem

Use Agentic Operations and the SRE Agent in Workflows to run a problem investigation. The agent builds on agentic root cause analysis, gathers evidence, and attaches its findings, root cause, and recommended fixes to the problem. Be sure to review the report before choosing your next step.

Choose the workflow action that matches the task:

• Run agent: Start the ready-made SRE investigation with preconfigured instructions and tools.
• Prompt agentic AI: Describe your task, choose the allowed tools, and get a structured result.
• Prompt generative AI: Generate a summary or suggested fix. Optionally ground it in Dynatrace documentation, Community, and Developer Portal content.

Start with templates for SRE Agent, Alert Reduction, Kubernetes Troubleshooting, Database Operations, or Infrastructure Optimization. Agents stay within configured tools and permissions. You control approval points, and the execution log records their steps. A recommended fix isn't an applied fix.

Review the SRE Agent's investigation findings and recommendations on the problem.
Use ready-made agents or configure AI actions for workflow tasks.

When you're investigating a dashboard or host yourself, ask Dynatrace Assist about the page open beside the chat. You don't need to describe it again. Assist tells you when it uses screen content and masks personal data with placeholders.

Causal root cause analysis also covers OpenTelemetry services. It combines related service anomalies into a single problem. Causal AI follows the Smartscape topology on Grail from monitored frontends into OpenTelemetry backends and from services to monitored databases.

Connect affected services and their dependencies to a single root cause.

Follow requests across queues without starting over

Now when you open a trace from a problem, you don't need to set the timeframe and filters again. Distributed Tracing carries them over with the problem ID and keeps that context visible in the chart. You keep it when you continue into Kubernetes, Logs, or Services, too.

For requests that cross a queue or serverless boundary, use automatic span-link resolution. When a consumer span contains OpenTelemetry span links, Distributed Tracing automatically loads upstream producer traces into the waterfall and highlights the linked spans. Open Linked traces to jump to a related trace at the connected span instead of searching for the producer by hand.

Keep the problem's timeframe and filters while inspecting its traces.
Follow linked traces across service and serverless boundaries in Distributed Tracing.

Find affected users with Apdex and location filters

Use Apdex scores on Grail to track satisfied, tolerating, and frustrated experiences across frontends and sessions. Then use location filters and frontend events to narrow the investigation, or open backend traces from Synthetic results.

• Experience scoring: Compare a 0-1 score per frontend and session. Classification uses user-action duration by default. Adjust classification and weighting with OpenPipeline rules when your reporting needs different thresholds.
• Location-specific investigations: Filter Experience Vitals, Error Inspector, and Users & Sessions by city, region, country, or continent to check whether a slowdown or error is local.
• Event naming and grouping: Choose frontend-reported names or define OpenPipeline rules so your event groups, dashboards, and alerts use your application's terms.
• Synthetic investigations: Open backend traces from HTTP monitor results. Configure maintenance windows to skip scheduled executions, mark runs during maintenance, and exclude those periods from availability calculations. Synthetic queries also exclude locations no longer monitored and use a shorter default time range.

Configure frontend event naming and grouping for downstream analysis.

Check network connections before debugging the application

Open Infrastructure & Operations to inspect host-to-host and process-to-process connections collected by OneAgent Network Connection Monitoring. In Kubernetes, inspect pod-to-pod, pod-to-service, and pod-to-process connections.

Check bandwidth, round-trip time, retransmissions, resets, and timeouts on the affected connection. These measurements give you network evidence before you change application code or hand the investigation to another team.

Inspect traffic and connection health between hosts and processes.
Inspect network connections between pods, services, and processes in Kubernetes.

Investigate Google Cloud alongside AWS and Azure

Open Google Cloud in Clouds to inspect metrics, logs, metadata, and topology alongside AWS and Azure. Use the ready-made dashboards and preconfigured health alerts to check resource health and investigate failures.

Onboard Google Cloud projects without deploying ActiveGate for metric polling. Google Cloud labels stay available across metrics, logs, and topology. Filter by the metadata you already use, then follow resource dependencies in Clouds and Smartscape.

The improved Clouds overview brings Google Cloud monitoring alongside AWS and Azure.

Share live dashboards without a Dynatrace account

Send customers, partners, or incident stakeholders a live dashboard link. They don't need a Dynatrace account. As the dashboard owner, you set the timeframe, refresh interval, and variable values they see.

Before sharing, review the data, ensure you want and are allowed to share, and set access restrictions. Use passcodes, expiration dates, and IP allowlists to control access to the link. You can also embed the dashboard in an iframe instead of exporting a static report.

Select Share externally to share a dashboard with viewers who do not have a Dynatrace account.
Configure an external sharing link before distributing it to viewers.

Performance, drilldowns, and navigation improvements

Return to recent views from the dock. Open Recents at the top of the dock and search for a screen by name. The list runs from newest to oldest. Each app also shows up to three recent visits; select one to return to the state you left. Recent views persist across browser sessions, and you can pin a view to the dock for easy access.

Inspect a node without leaving Smartscape. Select a node to open the updated details panel with its properties, key metrics, and active or resolved problems. Copy a property value, use it to filter the topology, or open a related Dynatrace app. Select a relationship to compare the two connected entities side by side.

Smartscape's Problem graph view shows affected entities and their relationships, with the details panel surfacing core fields, primary tags, and Dynatrace fields for the selected node.

See the cloud infrastructure behind a service. Open Service Map in Services to see the cloud compute a service runs on and its dependencies. Filter by cloud provider, region, or account, inspect the service's vertical topology, or open the underlying entities in Clouds. The service details panel also has a dedicated infrastructure tab, so you can check where the service runs without switching apps.

The Service Map shows the cloud infrastructure a service runs on, with drilldowns to Clouds and a vertical topology view.

Filter infrastructure by primary fields and tags. In Infrastructure & Operations, Kubernetes, Clouds, and Databases, filter by primary Grail fields such as Kubernetes namespaces, clusters, and cloud accounts. Use primary tags to narrow the view by team ownership or organizational boundaries. You'll find primary tags in a dedicated list-view column and in ready-made dashboard filters across all four apps.

These filters work alongside I&O's existing cross-app navigation, covered in August's Radar. Open related data in Logs, Clouds, Kubernetes, or Services with your filters, timeframe, and segments carried across.

The I&O row-action menu links to topology, traces, services, and more - with context carried into the target app.

Read dashboards and notebooks on mobile. Open a dashboard or notebook with more room for the data: the header moves out of the way, controls sit in a bottom sheet, and tile content no longer captures your scroll. Choose whether the detail panel overlays the dashboard or pushes it aside.

Find what's driving usage and cost. Use the ready-made dashboards in your environment to inspect Real User Monitoring and Agentic AI consumption alongside the existing Metrics usage view. The Agentic AI dashboard includes AI Units, Standard Function Calls, and AI-generated query consumption. Check all three rather than treating AI Units as the whole cost.

Filters and attribution vary by dashboard. For a Metrics spike, break usage down by source or metric prefix. For Agentic AI, inspect the users, workflows, caller types, and billing units behind the consumption.

Real User Monitoring	Metrics	Agentic AI Costs

Why this matters to practitioners

These updates help you spend less time piecing together an investigation: agent findings are added to the problem, your filters carry over when you open a trace, and you can follow linked requests across queues. Use experience scores to find affected frontends and sessions, and connection metrics to check the network before changing application code. Investigate Google Cloud resources alongside AWS and Azure, then share a live dashboard when someone else needs the same view. Check your environment's availability and the documented prerequisites before adopting an update.

Try the updates on our Release Radar launchpad on the Dynatrace Playground.