diff --git a/test/e2e-tests.sh b/test/e2e-tests.sh
index 32bb43b2..8d5bed88 100755
--- a/test/e2e-tests.sh
+++ b/test/e2e-tests.sh
@@ -15,8 +15,7 @@
 # limitations under the License.
 
 # This script runs the end-to-end tests against the build controller
-# built from source.
-# It is started by prow for each PR.
+# built from source. It is started by prow for each PR.
 # For convenience, it can also be executed manually.
 
 # If you already have the *_OVERRIDE environment variables set, call
@@ -26,54 +25,18 @@
 # Calling this script without arguments will create a new cluster in
 # project $PROJECT_ID, start the controller, run the tests and delete
 # the cluster.
-# $KO_DOCKER_REPO must point to a valid writable docker repo.
 
-source "$(dirname $(readlink -f ${BASH_SOURCE}))/library.sh"
-
-# Test cluster parameters and location of generated test images
-readonly E2E_CLUSTER_NAME=build-e2e-cluster${BUILD_NUMBER}
-readonly E2E_NETWORK_NAME=build-e2e-net${BUILD_NUMBER}
-readonly E2E_CLUSTER_ZONE=us-central1-a
-readonly E2E_CLUSTER_NODES=2
-readonly E2E_CLUSTER_MACHINE=n1-standard-2
-readonly TEST_RESULT_FILE=/tmp/build-e2e-result
-
-# This script.
-readonly SCRIPT_CANONICAL_PATH="$(readlink -f ${BASH_SOURCE})"
+# Load github.com/knative/test-infra/images/prow-tests/scripts/e2e-tests.sh
+[ -f /workspace/e2e-tests.sh ] \
+  && source /workspace/e2e-tests.sh \
+  || eval "$(docker run --entrypoint sh gcr.io/knative-tests/test-infra/prow-tests -c 'cat e2e-tests.sh')"
+[ -v KNATIVE_TEST_INFRA ] || exit 1
 
 # Helper functions.
 
 function teardown() {
-  header "Tearing down test environment"
-  # Free resources in GCP project.
-  if (( ! USING_EXISTING_CLUSTER )); then
-    ko delete --ignore-not-found=true -R -f test/
-    ko delete --ignore-not-found=true -f config/
-  fi
-
-  # Delete images when using prow.
-  if (( IS_PROW )); then
-    echo "Images in ${KO_DOCKER_REPO}:"
-    gcloud container images list --repository=${KO_DOCKER_REPO}
-    delete_gcr_images ${KO_DOCKER_REPO}
-  else
-    restore_override_vars
-  fi
-}
-
-function exit_if_test_failed() {
-  [[ $? -eq 0 ]] && return 0
-  echo "***************************************"
-  echo "***           TEST FAILED           ***"
-  echo "***    Start of information dump    ***"
-  echo "***************************************"
-  echo ">>> All resources:"
-  kubectl get all --all-namespaces
-  echo "***************************************"
-  echo "***           TEST FAILED           ***"
-  echo "***     End of information dump     ***"
-  echo "***************************************"
-  exit 1
+  ko delete --ignore-not-found=true -R -f test/
+  ko delete --ignore-not-found=true -f config/
 }
 
 function abort_test() {
@@ -81,98 +44,25 @@ function abort_test() {
   # If formatting fails for any reason, use yaml as a fall back.
   kubectl get builds -o=custom-columns-file=./test/columns.txt || \
     kubectl get builds -oyaml
-  false  # Force exit
-  exit_if_test_failed
+  fail_test
 }
 
 # Script entry point.
 
-cd ${BUILD_ROOT_DIR}
-
-# Show help if bad arguments are passed.
-if [[ -n $1 && $1 != "--run-tests" ]]; then
-  echo "usage: $0 [--run-tests]"
-  exit 1
-fi
-
-# No argument provided, create the test cluster.
-
-if [[ -z $1 ]]; then
-  header "Creating test cluster"
-  # Smallest cluster required to run the end-to-end-tests
-  CLUSTER_CREATION_ARGS=(
-    --gke-create-args="--enable-autoscaling --min-nodes=1 --max-nodes=${E2E_CLUSTER_NODES} --scopes=cloud-platform"
-    --gke-shape={\"default\":{\"Nodes\":${E2E_CLUSTER_NODES}\,\"MachineType\":\"${E2E_CLUSTER_MACHINE}\"}}
-    --provider=gke
-    --deployment=gke
-    --gcp-node-image=cos
-    --cluster="${E2E_CLUSTER_NAME}"
-    --gcp-zone="${E2E_CLUSTER_ZONE}"
-    --gcp-network="${E2E_NETWORK_NAME}"
-    --gke-environment=prod
-  )
-  if (( ! IS_PROW )); then
-    CLUSTER_CREATION_ARGS+=(--gcp-project=${PROJECT_ID:?"PROJECT_ID must be set to the GCP project where the tests are run."})
-  fi
-  # SSH keys are not used, but kubetest checks for their existence.
-  # Touch them so if they don't exist, empty files are create to satisfy the check.
-  touch $HOME/.ssh/google_compute_engine.pub
-  touch $HOME/.ssh/google_compute_engine
-  # Clear user and cluster variables, so they'll be set to the test cluster.
-  # KO_DOCKER_REPO is not touched because when running locally it must
-  # be a writeable docker repo.
-  export K8S_CLUSTER_OVERRIDE=
-  # Assume test failed (see more details at the end of this script).
-  echo -n "1"> ${TEST_RESULT_FILE}
-  kubetest "${CLUSTER_CREATION_ARGS[@]}" \
-    --up \
-    --down \
-    --extract "v${BUILD_GKE_VERSION}" \
-    --test-cmd "${SCRIPT_CANONICAL_PATH}" \
-    --test-cmd-args --run-tests
-  result="$(cat ${TEST_RESULT_FILE})"
-  echo "Test result code is $result"
-  exit $result
-fi
-
-# --run-tests passed as first argument, run the tests.
-
-# Set the required variables if necessary.
-
-if [[ -z ${K8S_USER_OVERRIDE} ]]; then
-  export K8S_USER_OVERRIDE=$(gcloud config get-value core/account)
-fi
-
-USING_EXISTING_CLUSTER=1
-if [[ -z ${K8S_CLUSTER_OVERRIDE} ]]; then
-  USING_EXISTING_CLUSTER=0
-  export K8S_CLUSTER_OVERRIDE=$(kubectl config current-context)
-  acquire_cluster_admin_role ${K8S_USER_OVERRIDE} ${E2E_CLUSTER_NAME} ${E2E_CLUSTER_ZONE}
-  # Make sure we're in the default namespace
-  kubectl config set-context $K8S_CLUSTER_OVERRIDE --namespace=default
-fi
-readonly USING_EXISTING_CLUSTER
-
-if [[ -z ${KO_DOCKER_REPO} ]]; then
-  export KO_DOCKER_REPO=gcr.io/$(gcloud config get-value project)/build-e2e-img
-fi
-
-# Build and start the controller.
-
-echo "- Cluster is ${K8S_CLUSTER_OVERRIDE}"
-echo "- User is ${K8S_USER_OVERRIDE}"
-echo "- Docker is ${KO_DOCKER_REPO}"
+initialize $@
+
+# Fail fast during setup.
+set -o errexit
+set -o pipefail
 
 header "Building and starting the controller"
-trap teardown EXIT
+export KO_DOCKER_REPO=${DOCKER_REPO_OVERRIDE}
+ko apply -f config/ || fail_test
 
-if (( USING_EXISTING_CLUSTER )); then
-  echo "Deleting any previous controller instance"
-  ko delete --ignore-not-found=true -f config/
-fi
+# Handle test failures ourselves, so we can dump useful info.
+set +o errexit
+set +o pipefail
 
-ko apply -f config/
-exit_if_test_failed
 # Make sure that are no builds or build templates in the current namespace.
 kubectl delete builds --all
 kubectl delete buildtemplates --all
@@ -181,14 +71,12 @@ kubectl delete buildtemplates --all
 
 header "Running tests"
 
-ko apply -R -f test/
-exit_if_test_failed
+ko apply -R -f test/ || fail_test
 
 # Wait for tests to finish.
 tests_finished=0
 for i in {1..60}; do
   finished="$(kubectl get builds --output=jsonpath='{.items[*].status.conditions[*].status}')"
-  echo $finished
   if [[ ! "$finished" == *"Unknown"* ]]; then
     tests_finished=1
     break
@@ -210,8 +98,8 @@ for expected_status in succeeded failed; do
       want=succeededfalse
       ;;
     *)
-      echo Invalid expected status $expected_status
-      exit 1
+      echo "Invalid expected status '${expected_status}'"
+      fail_test
   esac
   for result in ${results}; do
     if [[ ! "${result,,}" == *"=${want}" ]]; then
@@ -222,13 +110,4 @@ for expected_status in succeeded failed; do
 done
 (( tests_passed )) || abort_test "ERROR: one or more tests failed"
 
-# kubetest teardown might fail and thus incorrectly report failure of the
-# script, even if the tests pass.
-# We store the real test result to return it later, ignoring any teardown
-# failure in kubetest.
-# TODO(adrcunha): Get rid of this workaround.
-echo -n "0"> ${TEST_RESULT_FILE}
-echo "**************************************"
-echo "***        ALL TESTS PASSED        ***"
-echo "**************************************"
-exit 0
+success