# Debugging your Analytics Builder workflow becomes much more seamless: Introducing inline debugging & live model logging

**URL:** <https://community.cumulocity.com/t/debugging-your-analytics-builder-workflow-becomes-much-more-seamless-introducing-inline-debugging-live-model-logging/15313>\
**Category:** Product news\
**Tags:** analytics, Streaming-Analytics-Apama, analytics-builder, streaming-analytics\
**Created:** [October 28, 2026, 3:26pm UTC](https://community.cumulocity.com/t/debugging-your-analytics-builder-workflow-becomes-much-more-seamless-introducing-inline-debugging-live-model-logging/15313 "2026-10-28T15:26:42Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Nick\_Van\_Damme1](https://dub1.discourse-cdn.com/flex005/user_avatar/community.cumulocity.com/nick_van_damme1/32/31952_2.png) [@Nick\_Van\_Damme1](https://community.cumulocity.com/u/Nick_Van_Damme1)\
**Post date:** [October 28, 2026, 3:26pm UTC](https://community.cumulocity.com/t/debugging-your-analytics-builder-workflow-becomes-much-more-seamless-introducing-inline-debugging-live-model-logging/15313/1 "2026-10-28T15:26:42Z")

</div>

In our previous guide, [Debugging Strategies for Cumulocity Analytics Builder](https://community.cumulocity.com/t/debugging-strategies-for-cumulocity-analytics-builder/14451), we explored various strategies for inspecting model logic. At the time, troubleshooting required adding temporary intermediate output blocks—such as Measurement or Event blocks—and toggling back and forth between the Analytics Builder canvas, Device Management or Cockpit to trace data movement. While effective, context switching slowed down development and required managing temporary debug data.

With the latest release of Streaming Analytics, Analytics Builder is no longer a black box. You can now inspect internal states, trace data flows, and resolve runtime errors in real time—all directly within the model editing canvas.

 ![Inline Debugging for Cumulocity Analytics Builder](https://europe1.discourse-cdn.com/flex005/uploads/cumulocity/original/3X/6/2/629ae6bcea5aecf820c8fd9d23113c5373de82f9.jpeg)

## What’s New: Streamlined Inline Debugging

We’ve combined several new tools into a single, seamless debugging workflow:

- **Logger Block (General Availability):** Moving from public preview to GA, the **Logger** block lets you “tap the wire” at any point in your model logic to inspect data values without generating persistent database records.

- **Inline Model Log Viewer:** A dedicated real-time log viewer built straight into the Analytics Builder canvas. Watch signal values update live as events process.

- **One-Click Canvas Activation:** Activate or deactivate models into production or test mode directly from inside the editor workspace.

> **Video Walkthrough**
> 
> _Watch the demonstration during the October 2026 What’s New webinar to see inline debugging and real-time trace inspection in action:_
> 
> [https://www.youtube.com/watch?v=DWx2\_PsrSR8](https://www.youtube.com/watch?v=DWx2_PsrSR8)

## Hands-On Walkthrough: Inline Debugging in Action

### Scenario 1: Tracing Signal Flows (Pump Clogging Monitor)

Consider a standard model monitoring differential pressure across a pump to detect clogging and trigger an alarm when a threshold is exceeded.

If an alarm fails to trigger:

1. Drag **Logger** blocks from the _Utility_ section onto the canvas and connect them to key wires (e.g., input values, calculated difference, threshold output).

2. Assign tag names to each logger (such as `input_val`, `diff_pressure`, `after_threshold`).

3. Click **Activate** directly inside the canvas.

4. Open the **Model Logs** drawer to watch incoming values flow from left to right in real time.

You can instantly confirm whether the differential pressure calculation is accurate and verify the exact point where `after_threshold` evaluates to `true`—without leaving the canvas.

### Scenario 2: Diagnosing Runtime Errors (Dynamic Threshold Mismatches)

Suppose you replace a static threshold block with an **Expression** block to compare differential pressure against a dynamic threshold stored in a Managed Object. Upon activation, the model throws a runtime error: `Cannot apply operator >`.

Using inline debugging:

1. Attach **Logger** blocks to both inputs of the Expression block.

2. Review the live log output:

3. The comparison fails because of data type incompatibility.

4. Add a conversion block/expression to parse the string into a `float`, re-activate, and verify in the inline logs that both inputs now match as floats.

Once verified, simply hide or remove the logger blocks and resume normal operation.

## Direct Access to Microservice Logs

Additionally, for issues that go beyond an individual model’s logic, you no longer need to switch to the Administration app to check container logs. Microservice log access is now integrated directly into the Streaming Analytics application, giving you full operational diagnostics in one place.

Note: also within this interface, you can decide to switch context and only view the logs of a specific model.

## Try It Out

Inline debugging is designed to keep you in your flow state while building and optimizing streaming analytics logic.

Give the new inline logger and log viewer a spin in your model canvas, and let us know your thoughts! If you have feedback or unique debugging workflows, share them in the comments below.