JMeter Correlation Recorder 2.5: An Approach to the New Plugin Version

Legacy Correlation involves setting a custom rule set based on flow knowledge and identifying dynamic values needing correlation.

This mode closely resembles the manual correlation that JMeter flow automators use. It involves identifying both extraction and replacement regular expressions and re-recording to verify the process is accurate.

Next, we share a table with some real examples of extraction and replacement regular expressions:

In the latest version of the plugin, the process of extraction and replacement in .JSON format is made easier, allowing for more efficient correlation of values in structured data.

While this was possible before by using regular expressions for that purpose, now the use of JSON-path extractors simplifies the capture of specific information straightforwardly.

Upon closer examination of correlation through legacy, it requires the creation of replacement regular expressions that may sometimes be the same, similar, or different from the extraction ones. This entails an initial time investment to set up a set of rules, in scenarios where they may not be reusable.

For instance, scenarios like the following could illustrate this point, among others:

• An application frequently alters its structure or how dynamic values are created.
• Dynamic values are unique to a specific test scenario.
• Configured rules are highly personalized and specific to a particular case.

In such scenarios, if we were to manually automate the workflow, we would likely need to record with JMeter more than once to thoroughly explore the details and nuances of the process. This approach can lead to the introduction of errors that often cause many inconveniences (especially in extensive workflows).

From our experience, even in situations where manual automation of the workflow might seem faster at first glance, it can be beneficial to explore the plugin and correlation through legacy. This is because it reduces the likelihood of making mistakes and can help us address some complex scenarios.

Moreover, its ability to correlate flows, where requests rarely fail but might when reproduced for multiple users, makes this option even more appealing.

Initially, dynamic value linking through legacy was the only mode the plugin operated in. But then it evolved, incorporating two others in order to reduce time and effort.

Now, we’ll outline new ways to proceed, which need a single recording to work. They uniquely offer correlation suggestions if requests fail.

To correlate using automatic comparison and variable detection, it’s necessary to reproduce the initially recorded flow, identify failing requests, and present suggestions to make them work.

Next, we share a screenshot with the generated suggestions ready to apply in an example flow:

After applying the suggestions, correlation happens automatically. The test plan can then be checked to confirm the necessary values are extracted and replaced, making the flow run without errors.

Using existing templates for correlation means applying predefined rule sets to fix failing requests with relevant suggestions.

Templates can be saved locally, shared by the community, or provided by BlazeMeter. BlazeMeter’s templates follow best practices, and you can structure others similarly by following the plugin’s instructions.

We share a screenshot of the available templates for correlating an example flow:

At first, BlazeMeter provided some templates for certain application types, encouraging the community to add their own and expand the choices.

These templates load from the cloud, with their accessibility largely hinging on the type of platform subscription.

The application process is quite similar to the previous mode. Then, it’s just a matter of verifying proper functioning.

In automatic modes, it’s crucial to apply only the required suggestions for requests to work and not all that may appear. This way, we avoid adding unnecessary weight to the flow and excessive memory consumption.