Automatic Process Group detection in Dynatrace: a practical fix for service sprawl

Problem: locating services in a sprawling Smartscape

In large environments the Dynatrace Smartscape view can contain dozens or hundreds of individual process entries. When a service is represented by a single process, finding the right tile to investigate an alert becomes a needle‑in‑a‑haystack task. The result is wasted time, missed context, and slower incident response.

Why the friction matters

Developers and operators often need to correlate metrics, logs, and traces across many similar processes. Without a logical container, each alert is tied to a single process, making it hard to see the health of an entire tier such as a Java application or a micro‑service group.

Thesis: Automatic Process Group detection resolves the issue

The Automatic Process Group (APG) feature automatically creates logical containers in Smartscape. It groups processes that share the same executable, host naming pattern, or process name pattern, eliminating the need for manual tagging. The outcome is a single dashboard tile that aggregates CPU, memory, and error metrics for an entire tier, improving alert relevance and operational efficiency.

How it works

APG examines three key attributes of each process:

• Executable path – processes launched from the same binary.
• Host name – processes running on the same server or container.
• Naming pattern – process names that match a regular expression (e.g., *‑app.jar).

When a process is detected, Dynatrace evaluates these attributes against existing groups. If a match is found, the process is added to the group; otherwise a new group is created.

Step‑by‑step example

Below is a concrete workflow for a micro‑services environment that runs dozens of Java services on Linux hosts.

1. Confirm OneAgent version. The APG feature requires OneAgent 1.210 or newer. Run the following command on a representative host (requires admin rights on the host):

   curl -s https://download.dynatrace.com/agent/releases/ | grep -i "1.210"

   If the output shows a version ≥ 1.210, you are eligible to enable the feature.

2. Enable APG in OneAgent settings. In the Dynatrace UI go to Settings → OneAgent → Advanced and toggle Automatic Process Group detection to On. No restart of the agent is required; the change takes effect at the next metric collection cycle.
3. Verify initial groups. Open Smartscape, locate the Process Groups tile, and look for a group named Java Application. The tile should list processes whose executable path contains java and whose host name matches the pattern of your Java hosts.
4. Create a dashboard tile. Add a Process Group tile for Java Application. The tile will display aggregated CPU % and memory MB values for all member processes, giving you a single view of the entire tier.
5. Fine‑tune exclusions (optional). If you notice that unrelated services (e.g., a batch job named cleanup‑job.jar) are being grouped, edit the exclusion rule under Settings → OneAgent → Process Group exclusions. Add a regex that matches the unwanted pattern, then save.

Verification checklist

• OneAgent version ≥ 1.210 (check via dtagent -v on the host).
• Smartscape shows a logical group containing multiple processes that share the same executable or naming pattern.
• Dashboard tile for the group reflects aggregated metrics rather than a single process.

Trade‑off and limitation

While APG dramatically reduces manual grouping effort, it can produce overly broad groups when services share a naming pattern. For example, a set of services whose JAR files all end with -service.jar may be merged into one group, causing alerts for one component to mask problems in another. Additionally, heavy auto‑grouping can increase backend query load, especially in high‑volume environments, which may manifest as slightly slower dashboard refreshes. Teams should review generated groups after the initial rollout and adjust exclusion rules to keep the view accurate.

Actionable close

To adopt Automatic Process Group detection responsibly:

1. Ensure every host runs OneAgent 1.210 or newer.
2. Enable the feature in the OneAgent advanced settings.
3. Validate that the automatically created groups contain the expected processes by inspecting the Smartscape Process Group tile.
4. Add dashboard tiles for the most important groups to monitor aggregated health metrics.
5. Periodically review exclusion rules to prevent false positives and to keep query load manageable.

By following these steps you can turn a chaotic Smartscape into a clear, actionable view of your Java tier, improving both detection speed and mean‑time‑to‑resolution.