Validate thresholds

Validate thresholds work

Adding thresholds to your script isn’t enough. You need to confirm they actually pass and fail as expected. In this milestone, you run the test twice: once to verify it passes under normal conditions, and once with a deliberately strict threshold to confirm it fails correctly.

To validate your thresholds, complete the following steps:

Run the test:
Bash
```
k6 run baseline.js
```
At the top of the summary output, look for the THRESHOLDS section and verify that all thresholds pass:
text
```
THRESHOLDS
  checks
    ✓ 'rate>0.99' rate=100.00%

  http_req_duration
    ✓ 'p(95)<500' p(95)=54.42ms

  http_req_failed
    ✓ 'rate<0.01' rate=0.00%
```
A ✓ next to each threshold means it passed. The rate>0.99 on the checks threshold means the check pass rate must be above 99%. This is not related to p99 latency. The exit code should be 0, which signals success.
Verify the exit code in your terminal:
Bash
```
echo $?
```
The output should be 0.
Now test that a threshold failure works correctly. Temporarily set a very strict latency threshold that your system can’t meet. Open baseline.js and change the http_req_duration threshold:
JavaScript
```
thresholds: {
  http_req_duration: ['p(95)<1'],
  http_req_failed: ['rate<0.01'],
  checks: ['rate>0.99'],
},
```
A p95 under 1 ms is impossible for any real HTTP request, so this threshold should fail.

Run the test again:

k6 run baseline.js

The THRESHOLDS section now shows a ✗ next to the failed threshold:

THRESHOLDS
  checks
    ✓ 'rate>0.99' rate=100.00%

  http_req_duration
    ✗ 'p(95)<1' p(95)=54.42ms

  http_req_failed
    ✓ 'rate<0.01' rate=0.00%

Verify the exit code is non-zero:
Bash
```
echo $?
```
The output should be 99, which is k6’s exit code for threshold failures. This non-zero exit code is what CI/CD systems use to detect a failed test.
Restore your original threshold value before continuing:
JavaScript
```
http_req_duration: ['p(95)<500'],
```

You’ve confirmed that your thresholds work in both directions: passing when metrics are within bounds and failing when they’re breached.

In the next milestone, you save your results to k6 Cloud for historical tracking and team visibility.

Add thresholds

page 8 of 10

Save and share

page 8 of 10