Is it possible to track https status codes for requests in k6 output?

Grafana k6 OSS Support
1

Hello!

Please, clarify if it’s possible to show output statistics which is grouped by http status codes for each request (200 OK, 40* errors, 50* errors) and how ?

For example:

I have a load test for one specific method which is grouped by group function.
I need to see how many requests for the method were executed correctly (200 OK) and how many errors produced - what type (40*, 50* ) if they were occured.

 █ Events

  ✓ status is 200 OK
  ✓ content-type is application/json

GetEvents..................: avg=6640.0475 min=6624.679 med=6643.218 max=6647.889 p(90)=6646.6551 p(95)=6647.27205
checks.....................: 100.00% ✓ 20   ✗ 0   
data_received..............: 97 kB   1.6 kB/s
data_sent..................: 10 kB   173 B/s
errors.....................: 0.00%   ✓ 0    ✗ 10  
group_duration.............: avg=7.14s     min=7.13s    med=7.15s    max=7.15s    p(90)=7.15s     p(95)=7.15s     
http_req_blocked...........: avg=494.26ms  min=491.84ms med=494.62ms max=496.55ms p(90)=496.21ms  p(95)=496.38ms  
http_req_connecting........: avg=45.31ms   min=40.64ms  med=47.47ms  max=48ms     p(90)=47.95ms   p(95)=47.97ms   
http_req_duration..........: avg=6.64s     min=6.62s    med=6.64s    max=6.64s    p(90)=6.64s     p(95)=6.64s     
http_req_receiving.........: avg=10.95ms   min=1.62ms   med=5.85ms   max=23.74ms  p(90)=23.16ms   p(95)=23.45ms   
http_req_sending...........: avg=1.45ms    min=220µs    med=312µs    max=3.24ms   p(90)=3.22ms    p(95)=3.23ms    
http_req_tls_handshaking...: avg=356.47ms  min=350.72ms med=356.17ms max=363.35ms p(90)=363.32ms  p(95)=363.33ms  
http_req_waiting...........: avg=6.62s     min=6.61s    med=6.62s    max=6.64s    p(90)=6.64s     p(95)=6.64s     
http_reqs..................: 10      0.166391/s
iteration_duration.........: avg=7.14s     min=7.13s    med=7.15s    max=7.15s    p(90)=7.15s     p(95)=7.15s     
iterations.................: 10      0.166391/s
vus........................: 10      min=10 max=10
vus_max....................: 10      min=10 max=10
1 Like
2

It is somewhat possible, but it requires a bit of a hack until Add explicit tracking and ignoring of metrics and sub-metrics · Issue #1321 · grafana/k6 · GitHub is implemented. There are some examples linked in this docs issue: Add example for sub-metrics by scenario (or other tags) in summary · Issue #205 · grafana/k6-docs · GitHub

Here’s how this can look like in your case:

import http from 'k6/http'

export let options = {
    vus: 5,
    iterations: 10,
    thresholds: {
        // Some dummy thresholds that are always going to pass.
        'http_req_duration{status:200}': ['max>=0'],
        'http_req_duration{status:403}': ['max>=0'],
        'http_req_duration{method:POST}': ['max>=0'],
    },
    'summaryTrendStats': ['min', 'med', 'avg', 'p(90)', 'p(95)', 'max', 'count'],
}

export default function () {
    http.get('https://httpbin.test.k6.io/status/200');
    http.get('https://httpbin.test.k6.io/anything');
    http.get('https://httpbin.test.k6.io/status/403');
    http.post('https://httpbin.test.k6.io/anything', 'some body');
}

will result in something like this:

...
     ✓ { method:POST }..............: min=123.86ms med=127.57ms avg=128.64ms p(90)=134.27ms p(95)=134.38ms max=134.5ms  count=10
     ✓ { status:200 }...............: min=121.81ms med=127.8ms  avg=128.43ms p(90)=133.77ms p(95)=134.35ms max=134.5ms  count=30
     ✓ { status:403 }...............: min=122.05ms med=128.74ms avg=129.24ms p(90)=134.14ms p(95)=135.2ms  max=136.26ms count=10
...

Notice that you currently can’t specify status ranges, only single status values in the sub-metrics.

2 Likes
3

Faced the same problem recently. If you only need count / distribution of status codes, then I worked it around with this

  check(res, {
    'status was 200': (r) => r.status == 200,
    'status was 422': (r) => r.status == 422,
    'status was 423': (r) => r.status == 422,
    'status was 500': (r) => r.status == 500,
    'no response': (r) => !r.status,
  })

It gives you an output like

     ✗ status was 200
      ↳  68% — ✓ 5103 / ✗ 2330
     ✗ status was 422
      ↳  0% — ✓ 0 / ✗ 7433
     ✗ status was 423
      ↳  0% — ✓ 0 / ✗ 7433
     ✗ status was 500
      ↳  0% — ✓ 0 / ✗ 7433
     ✗ no response
      ↳  31% — ✓ 2330 / ✗ 5103

As it’s just a JS you can also match by wildcard, i.e. 4xx / 5xx

2 Likes