Include kubernetes cron example?

[zalegrala]

This project might make an interesting "verification" step when standing up a new Tempo cluster. Currently, we document a few ways of getting data into the cluster which requires users to have already instrumented something and to be ready to start pushing traces. Could this replace the deprecated synthetic load generator?

Here is perhaps an easy way for kubernetes users to get started using k6, taken from a lab environment.

apiVersion: v1
data:
  param.js: |
    import { sleep } from 'k6'
    import tracing from 'k6/x/tracing'
    import { randomIntBetween } from 'https://jslib.k6.io/k6-utils/1.2.0/index.js'

    export let options = {
      vus: 1,
      duration: '1m',
    }

    const client = new tracing.Client({
      endpoint: 'tempo.trace.svc.cluster.znet:4317',
      exporter: tracing.EXPORTER_OTLP,
      insecure: true,
      headers: { 'X-Scope-OrgID': 'k6' },
    })

    export default function () {
      let pushSizeTraces = randomIntBetween(2, 3)
      let pushSizeSpans = 0
      let t = []
      for (let i = 0; i < pushSizeTraces; i++) {
        let c = randomIntBetween(5, 35)
        pushSizeSpans += c

        t.push({
          random_service_name: false,
          spans: {
            count: c,
            size: randomIntBetween(300, 30000),
            random_name: true,
            fixed_attrs: {
              test: 'test',
            },
          },
        })
      }

      let gen = new tracing.ParameterizedGenerator(t)
      let traces = gen.traces()
      client.push(traces)

      console.log(
        `Pushed ${pushSizeSpans} spans from ${pushSizeTraces} different traces. Here is a random traceID: ${
          t[Math.floor(Math.random() * t.length)].id
        }`
      )
      sleep(1)
    }

    export function teardown() {
      client.shutdown()
    }
kind: ConfigMap
metadata:
  name: k6-tracing-config
  namespace: trace
---
apiVersion: batch/v1
kind: CronJob
metadata:
  name: k6-tracing
  namespace: trace
spec:
  jobTemplate:
    spec:
      backoffLimit: 1
      template:
        metadata:
          labels:
            name: k6-tracing
        spec:
          containers:
          - args:
            - run
            - /config/param.js
            image: zalegrala/xk6-client-tracing:latest
            imagePullPolicy: Always
            name: k6-tracing
            ports:
            - containerPort: 6565
              name: none
            volumeMounts:
            - mountPath: /config
              name: k6-tracing-config
          restartPolicy: Never
          volumes:
          - configMap:
              name: k6-tracing-config
            name: k6-tracing-config
      ttlSecondsAfterFinished: 604800
  schedule: '*/10 * * * *'

The above CronJob will run k6 every 10 minutes for 1 minute. Is this worth working into this project, and perhaps writing a bit of documentation for users to verify a new cluster? Perhaps there is value in keeping a job around even if you have a running cluster.