In SteelCentral NetSensor, when setting up a synthetic test with External Script, for example a .cmd, the Response Time recorded (and displayed in graphs) is the total duration for each successful execution (success is when script exit code is 0, failure when different than 0)
In some cases it would be great to have more granularity and be able to control the response time measurement from the inside of the External Script that runs on the Synthetic source. For example, when the External Script is a complex script that prepares lot of things on the station before actually running the particular "action" or "scenario" we need to measure, there are some unwanted long or random delays that add to the total script execution duration.
Here, the suggestion is to extend NetSensor "external script" with a feature to return the Response Time measured by the script itself.
- Idea 1: Use the script exit code as the Response Time value (millisecond)
This makes new specifications for the script exit code (numerical value): zero or a positive codes mean synthetic test execution success. The value is the test Response Time measured and returned by the script, that NetSensor must take as the synthetic test Response Time (graphs, monitoring, alerting). The unit is millisecond.
Negative codes are dedicated to script errors and keep raising synthetic test execution failure to NetSensor.
The feature might be embedded in the Advanced section of External Script configuration (see screenshot) adding a dedicated checkbox to enable or disable.
- Idea 2: Provide an API useable by External Script to return the "measured response time" (API call with a kind of performance beacon, REST, SteelScript,...)