## Reference

### Design details

Broadly, `istio-cni` accomplishes ambient redirection by instructing ztunnel to set up sockets within the application pod network namespace, where:

- one end of the socket is in the application pod
- and the other end is in ztunnel's pod

and setting up iptables rules to funnel traffic thru that socket "tube" to ztunnel and back.

fi

export RELEASE=${1}

docker-compose -f docker-compose-site1.yaml up -d
docker-compose -f docker-compose-site2.yaml up -d

./mc ready site1/
./mc ready site2/

for i in $(seq 1 10); do
	# mc admin heal -r --remove when used against a LB endpoint
	# behaves flaky, let this run 10 times before giving up
	./mc admin heal -r --remove --json site1/ 2>&1 >/dev/null

        with:
          access_token: ${{ github.token }}
      - name: 'Check out repository'
        uses: actions/checkout@0ad4b8fadaa221de15dcec353f45205ec38ea70b # v4.1.4
      - name: 'Set up JDK ${{ matrix.java }}'
        uses: actions/setup-java@99b8673ff64fbf99d8d325f52d9a5bdedb8483e9 # v4.2.1

        with:
          java-version: ${{ matrix.java }}
          distribution: 'zulu'
          cache: 'maven'

 * guaranteed to execute. There is no guarantee, however, that listeners will be executed in the
 * order that they are added.
 *
 * <p>Exceptions thrown by a listener will be propagated up to the executor. Any exception thrown
 * during {@code Executor.execute} (e.g., a {@code RejectedExecutionException} or an exception
 * thrown by {@linkplain MoreExecutors#directExecutor direct execution}) will be caught and logged.
 *

    if (nextPlanIndex >= plans.size && autoGeneratePlans) addPlan()

    require(nextPlanIndex < plans.size) {
      "not enough plans! call addPlan() or set autoGeneratePlans=true in the test to set this up"
    }
    val result = plans[nextPlanIndex++]
    events += "take plan ${result.id}"

    if (result.yieldBeforePlanReturns) {
      taskFaker.yield()
    }

    val planningThrowable = result.planningThrowable

This platform is agnostic to what exactly the purpose of the work is.
It might be creating an application, setting up development environments, orchestrating deployments, running simulations, etc.

This platform does not provide special support for a particular kind of automation. This is the responsibility of other platforms.

It is made up of 3 architecture modules:

   *
   * We get plan0 and plan1 from the route planner.
   * We get plan2 as a follow-up to plan1, typically retry the same IP but different TLS.
   * We get plan3 as a retry of plan0, which was canceled when it lost the race.
   *
   * This test confirms that we prefer to do the TLS follow-up (plan2) before the TCP retry (plan3).
   * It also confirms we enforce the 250 ms delay in each race.
   */
  @Test

        assertEquals(
            listOf(
                "KILL_LEAKED_PROCESSES_FROM_PREVIOUS_BUILDS",
                "GRADLE_RUNNER",
                "KILL_ALL_GRADLE_PROCESSES",
                "CLEAN_UP_GIT_UNTRACKED_FILES_AND_DIRECTORIES",
                "GRADLE_RETRY_RUNNER",
                "KILL_PROCESSES_STARTED_BY_GRADLE",
                "CHECK_CLEAN_M2_ANDROID_USER_HOME"
            ),

All the **logic** about how to register the URLs for webhooks and the code to actually send those requests is up to you. You write it however you want to in **your own code**.

## Documenting webhooks with **FastAPI** and OpenAPI

    // or we can show it graphically as:
    //
    // p1 ---> p0 --> super model
    //
    // ----------------------------------------------------------------------

    /**
     * How the test project is set up:
     *
     * 1. dependencyManagement lists dependencies on a & b,
     *    with an exclusion on c in b.
     * 2. the child project lists a dependency on project a only