The Failure Criteria feature allows you to set the pass or fail criteria of your test for various metrics, such as response times, errors, hits per second, test duration, and so on.
- Enable and Define Failure Criteria
- Define Failure Criteria Using the Baseline
- Fail Tests against the Baseline when Running from Taurus
- Baseline-based Taurus failure criteria in BlazeMeter UI
Enable and Define Failure Criteria
To enable failure criteria, follow these steps:
- In the Performance tab, click Create Test, Performance Test.
- Scroll down to the Failure Criteria section and toggle the button ON.
The following parameters are available:
Specify if you want to use this rule on a particular label from your script. It is set to ALL (all labels) by default.
- Key Performance Indicator
Select the specific metric you want to apply a rule for. Expand the drop-down menu and review available metrics to monitor.
To see the binary comparison operators for this rule, expand the drop-down menu. Operators include Less than, Greater than, Equal to, and Not Equal to.
The numeric value you want this rule to apply to.
- Stop Test
If this box is checked, the test will stop immediately when that criteria fails. Otherwise, it will continue running uninterrupted.
- Delete Failure Criteria
Click the trash bin icon to delete the criteria.
- Ignore failure criteria during ramp-up (Advanced configuration drop-down)
Checking this box causes the test to ignore any failures that occur during ramp-up, so that criteria are only evaluated after ramp-up ends.
If implemented, the pass / failure results can be seen via the workspace dashboard, making it easier to monitor your testing over time. The dashboard is accessed by selecting the Workspace of interest from the Workspace menu drop-down list.
Note: For Taurus tests (tests that use a YAML configuration file), we are translating the majority of Taurus pass/fail capabilities into Blazemeter Failure Criteria, so that when a YAML script is uploaded to Blazemeter, the pass/fail module in the script will automatically appear in the test UI. You can also execute a test from Taurus with cloud provisioning, and the pass/fail module will be recognized by BlazeMeter and displayed in the report.
Define Failure Criteria Using the Baseline
One of the most useful things about having a baseline to compare with, and fail the test in comparison to it (baseline-based failure criteria), is to have more accurate failures when automating your tests.
When you set a report as a baseline, you can also set the failure criteria against the baseline. If performance degrades compared to the baseline in a subsequent test, the test will fail automatically.
Follow these steps:
- Go to the Performance tab, click Tests and select a test from the drop down list.
The test reports window opens.
- Click the Configuration tab and scroll down to the Failure Criteria section.
- Toggle the button for Failure Criteria on.
- Check the box for Use from baseline. This checkbox is available only if there is a baseline selected for the test.
- You can:
- Set the threshold for each failure criteria from the selected baseline.
- Define an offset from baseline you want to tolerate, so that minor deviation from baseline will not be defined as a failure.
If you use Baseline to define the threshold, the threshold will be visible after save.
Subsequent tests are compared against the baseline with these failure criteria. If performance degrades compared to the baseline and the defined offset, the test will fail automatically.
Fail Tests against the Baseline when Running from Taurus
In this section we will focus on defining baseline-based failure criteria in Taurus.
The failure criteria using the Baseline can be configured in the YAML and "translated" into BlazeMeter failure criteria.
Note: Failing tests from Taurus in comparison to the baseline is only possible in case the test is already defined in BlazeMeter (not for new tests created via Taurus), and in case the test has a baseline defined in BlazeMeter (you can’t set a baseline for a test while running it in Taurus).
You can define the failure criteria in the YAML file. This can be done when:
- Running YAML test in BlazeMeter.
- Running existing Taurus tests from Command line in the cloud (in BlazeMeter) and not locally.
For each pass/fail criteria, you can define if it is taken from baseline and with which offset.
- module: passfail
- avg-rt>baseline, stop as failed
- avg-rt>baseline+5%, stop as failed
- hits<baseline-5%, stop as failed
The above example uses baseline as a threshold for the criteria.
The format is the following:
[KPI] [Condition] baseline [Offset]
- KPI: the KPI that will be compared
- Condition: the comparison operator, one of >, <, >=, <=, =, == (same as =).
- Baseline: the phrase “baseline” indicates that the value to compare with (the threshold) is taken from the baseline.
- Offset: the percent of deviation from baseline that should be tolerable and not fail the test
The following validations apply on the Condition and Offset:
- If the condition is > or >=, the offset must be defined, with a "+" sign.
- If the condition is < or <=, the offset sign must be defined, with a "-" sign.
- If the condition is = or == , offset must not be defined.
In case validation failed or syntax is incorrect, the pass/fail criteria will not be recognized and enforced by BlazeMeter.
Baseline-Based Taurus Failure Criteria in BlazeMeter UI
Like other Taurus pass/fail criteria, baseline-based criteria in YAML are translated to BlazeMeter failure criteria and appear in the test configuration UI. “Stop test” in Taurus is not supported in BlazeMeter.
- Key Performance Indicator: the KPI defined in the pass/fail criteria
- Condition: the comparison operator in the pass/fail criteria will be translated to the parallel phrase, e.g. “>=” will be translated to "Greater than or equal to".
- Use from baseline? : true
- Offset = the percent defined in the pass/fail without the +/- sign. The +/- sign will be used to add or reduce the offset percent from the baseline when calculating the Threshold.
- Threshold: the value is calculated from the baseline and offset. For example, if average response time in baseline is 100 ms and the criteria had a “+5%” offset, the threshold would be baseline+5% which is 105 ms.
Jump to next section: