diff --git a/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/bootstrap-in-place.sh b/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/bootstrap-in-place.sh
index 2661c40edb4..00e875c0fd3 100755
--- a/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/bootstrap-in-place.sh
+++ b/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/bootstrap-in-place.sh
@@ -12,6 +12,7 @@ bootkube_podman_run() {
 }
 
 if [ ! -f stop-etcd.done ]; then
+  record_service_stage_start "stop-etcd"
   echo "Stop etcd static pod by moving the manifest"
   mv /etc/kubernetes/manifests/etcd-member-pod.yaml /etc/kubernetes || echo "already moved etcd-member-pod.yaml"
 
@@ -21,9 +22,11 @@ if [ ! -f stop-etcd.done ]; then
   done
 
   touch stop-etcd.done
+  record_service_stage_success
 fi
 
 if [ ! -f master-ignition.done ]; then
+  record_service_stage_start "master-ignition"
   echo "Creating master ignition and writing it to disk"
   # Get the master ignition from MCS
   curl --header 'Accept:application/vnd.coreos.ignition+json;version=3.2.0' \
@@ -50,4 +53,5 @@ if [ ! -f master-ignition.done ]; then
     --output /assets/master.ign
 
   touch master-ignition.done
+  record_service_stage_success
 fi
diff --git a/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/install-to-disk.sh.template b/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/install-to-disk.sh.template
index 7d6f8c88ea9..6996e593add 100644
--- a/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/install-to-disk.sh.template
+++ b/data/data/bootstrap/bootstrap-in-place/files/usr/local/bin/install-to-disk.sh.template
@@ -3,20 +3,28 @@ set -euoE pipefail ## -E option will cause functions to inherit trap
 
 # This script is executed by install-to-disk service when installing single node with bootstrap in place
 
+. /usr/local/bin/bootstrap-service-record.sh
+
+record_service_stage_start "wait-for-bootkube"
 echo "Waiting for /opt/openshift/.bootkube.done"
 until [ -f /opt/openshift/.bootkube.done ]; do
   sleep 5
 done
+record_service_stage_success
 
 if [ ! -f coreos-installer.done ]; then
+  record_service_stage_start "coreos-installer"
   # Write image + ignition to disk
   echo "Executing coreos-installer with the following options: install -i /opt/openshift/master.ign {{.BootstrapInPlace.InstallationDisk}}"
   coreos-installer install -i /opt/openshift/master.ign {{.BootstrapInPlace.InstallationDisk}}
 
   touch coreos-installer.done
+  record_service_stage_success
 fi
 
+record_service_stage_start "reboot"
 echo "Going to reboot"
 shutdown -r +1 "Bootstrap completed, server is going to reboot."
 touch /opt/openshift/.install-to-disk.done
 echo "Done"
+record_service_stage_success
diff --git a/data/data/bootstrap/files/usr/local/bin/approve-csr.sh b/data/data/bootstrap/files/usr/local/bin/approve-csr.sh
index f0018e7e476..2372d39f4b3 100644
--- a/data/data/bootstrap/files/usr/local/bin/approve-csr.sh
+++ b/data/data/bootstrap/files/usr/local/bin/approve-csr.sh
@@ -1,5 +1,8 @@
 #!/usr/bin/env bash
 
+# shellcheck disable=SC1091  # using path on bootstrap machine
+. /usr/local/bin/bootstrap-service-record.sh
+
 KUBECONFIG="${1}"
 
 echo "Approving all CSR requests until bootstrapping is complete..."
diff --git a/data/data/bootstrap/files/usr/local/bin/bootkube.sh.template b/data/data/bootstrap/files/usr/local/bin/bootkube.sh.template
index cb6df518e1b..84c707fe304 100755
--- a/data/data/bootstrap/files/usr/local/bin/bootkube.sh.template
+++ b/data/data/bootstrap/files/usr/local/bin/bootkube.sh.template
@@ -1,6 +1,8 @@
 #!/usr/bin/env bash
 set -euoE pipefail ## -E option will cause functions to inherit trap
 
+. /usr/local/bin/bootstrap-service-record.sh
+
 . /usr/local/bin/release-image.sh
 
 mkdir --parents /etc/kubernetes/{manifests,bootstrap-configs,bootstrap-manifests}
@@ -68,13 +70,16 @@ mkdir --parents ./{bootstrap-manifests,manifests}
 
 if [ ! -f openshift-manifests.done ]
 then
+    record_service_stage_start "openshift-manifests"
 	echo "Moving OpenShift manifests in with the rest of them"
 	cp openshift/* manifests/
 	touch openshift-manifests.done
+    record_service_stage_success
 fi
 
 if [ ! -f cvo-bootstrap.done ]
 then
+    record_service_stage_start "cvo-bootstrap"
 	echo "Rendering Cluster Version Operator Manifests..."
 
 	rm --recursive --force cvo-bootstrap
@@ -95,11 +100,13 @@ then
 	cp auth/kubeconfig-loopback /etc/kubernetes/kubeconfig
 
 	touch cvo-bootstrap.done
+    record_service_stage_success
 fi
 
 ETCD_ENDPOINTS=https://localhost:2379
 if [ ! -f etcd-bootstrap.done ]
 then
+    record_service_stage_start "etcd-bootstrap"
 	echo "Rendering CEO Manifests..."
 	bootkube_podman_run \
 		--volume "$PWD:/assets:z" \
@@ -131,10 +138,12 @@ then
 	fi
 
 	touch etcd-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f config-bootstrap.done ]
 then
+    record_service_stage_start "config-bootstrap"
 	echo "Rendering cluster config manifests..."
 
 	rm --recursive --force config-bootstrap
@@ -158,10 +167,12 @@ then
 	cp config-bootstrap/manifests/* manifests/
 
 	touch config-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f kube-apiserver-bootstrap.done ]
 then
+    record_service_stage_start "kube-apiserver-bootstrap"
 	echo "Rendering Kubernetes API server core manifests..."
 
 	rm --recursive --force kube-apiserver-bootstrap
@@ -185,10 +196,12 @@ then
 	cp kube-apiserver-bootstrap/manifests/* manifests/
 
 	touch kube-apiserver-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f kube-controller-manager-bootstrap.done ]
 then
+    record_service_stage_start "kube-controller-manager-bootstrap"
 	echo "Rendering Kubernetes Controller Manager core manifests..."
 
 	rm --recursive --force kube-controller-manager-bootstrap
@@ -211,10 +224,12 @@ then
 	cp kube-controller-manager-bootstrap/manifests/* manifests/
 
 	touch kube-controller-manager-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f kube-scheduler-bootstrap.done ]
 then
+    record_service_stage_start "kube-scheduler-bootstrap"
 	echo "Rendering Kubernetes Scheduler core manifests..."
 
 	rm --recursive --force kube-scheduler-bootstrap
@@ -233,10 +248,12 @@ then
 	cp kube-scheduler-bootstrap/manifests/* manifests/
 
 	touch kube-scheduler-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f ingress-operator-bootstrap.done ]
 then
+    record_service_stage_start "ingress-operator-bootstrap"
 	echo "Rendering Ingress Operator core manifests..."
 
 	rm --recursive --force ingress-operator-bootstrap
@@ -251,10 +268,12 @@ then
 	cp ingress-operator-manifests/* manifests/
 
 	touch ingress-operator-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f mco-bootstrap.done ]
 then
+    record_service_stage_start "mco-bootstrap"
 	echo "Rendering MCO manifests..."
 
 	rm --recursive --force mco-bootstrap
@@ -323,10 +342,12 @@ then
 	cp tls/machine-config-server.key /etc/ssl/mcs/tls.key
 
 	touch mco-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ ! -f cco-bootstrap.done ]
 then
+    record_service_stage_start "cco-bootstrap"
 	echo "Rendering CCO manifests..."
 
 	rm --recursive --force cco-bootstrap
@@ -349,13 +370,16 @@ then
 	fi
 
 	touch cco-bootstrap.done
+    record_service_stage_success
 fi
 
 # in case of single node, if we removed etcd, there is no point to wait for it on restart
 if [ ! -f stop-etcd.done ]
 then
+    record_service_stage_start "wait-for-etcd"
 	# Wait for the etcd cluster to come up.
 	wait_for_etcd_cluster
+    record_service_stage_success
 fi
 
 REQUIRED_PODS="openshift-kube-apiserver/kube-apiserver,openshift-kube-scheduler/openshift-kube-scheduler,openshift-kube-controller-manager/kube-controller-manager,openshift-cluster-version/cluster-version-operator"
@@ -368,6 +392,7 @@ echo "Starting cluster-bootstrap..."
 
 if [ ! -f cb-bootstrap.done ]
 then
+    record_service_stage_start "cb-bootstrap"
     bootkube_podman_run \
         --rm \
         --volume "$PWD:/assets:z" \
@@ -376,6 +401,7 @@ then
         start --tear-down-early=false --asset-dir=/assets --required-pods="${REQUIRED_PODS}"
 
     touch cb-bootstrap.done
+    record_service_stage_success
 fi
 
 if [ "$BOOTSTRAP_INPLACE" = true ]
@@ -386,6 +412,7 @@ else
 
     if [ ! -z "$CLUSTER_ETCD_OPERATOR_IMAGE" ]
     then
+        record_service_stage_start "wait-for-ceo"
         echo "Waiting for CEO to finish..."
         bootkube_podman_run \
             --volume "$PWD:/assets:z" \
@@ -393,8 +420,10 @@ else
             /usr/bin/cluster-etcd-operator \
             wait-for-ceo \
             --kubeconfig /assets/auth/kubeconfig
+        record_service_stage_success
     fi
 fi
+
 # Workaround for https://github.com/opencontainers/runc/pull/1807
 touch /opt/openshift/.bootkube.done
 echo "bootkube.service complete"
diff --git a/data/data/bootstrap/files/usr/local/bin/bootstrap-service-record.sh b/data/data/bootstrap/files/usr/local/bin/bootstrap-service-record.sh
new file mode 100644
index 00000000000..66cb04abed5
--- /dev/null
+++ b/data/data/bootstrap/files/usr/local/bin/bootstrap-service-record.sh
@@ -0,0 +1,158 @@
+#!/usr/bin/env bash
+# This library provides a helper functions for recording when a service
+# and its stages start and end.
+
+###
+# When running as a pre- or post-command, set PRE_COMMAND or POST_COMMAND, respectively.
+# These must be set *prior* to sourcing this script.
+# PRE_COMMAND is the name identifying the pre-command being run.
+# POST_COMMAND is the name identifying the post-command being run.
+
+# SERVICE_RECORDS_DIR is the directory under which service records will be stored.
+SERVICE_RECORDS_DIR="${SERVICE_RECORDS_DIR:-/var/log/openshift/}"
+# SYSTEMD_UNIT_NAME is the name of the systemd unit for the service
+SYSTEMD_UNIT_NAME="$(ps -o unit= $$)"
+# SERVICE_NAME is the name of the service
+SERVICE_NAME="${SERVICE_NAME:-${SYSTEMD_UNIT_NAME%.service}}"
+
+# add_service_record_entry adds a record entry to the service records file.
+#   PHASE - phase being recorded; one of "service start", "service end", "stage start", "stage end", "pre-command start",
+#           "pre-command end", "post-command start", "post-command end"
+#   RESULT - result of the action
+#   STAGE (optional) - stage of the service
+#   PRE_COMMAND (optional) - name of the pre-command
+#   POST_COMMAND (optional) - name of the post-command
+#   ERROR_LINE (optional) - line where the error occurred
+#   ERROR_MESSAGE (optional) - message for the error
+add_service_record_entry() {
+  local FILENAME="${SERVICE_RECORDS_DIR}/${SERVICE_NAME}.json"
+  mkdir --parents "$(dirname "${FILENAME}")"
+  # Append the new entry to the existing array in the file.
+  # If the file does not already exist, start with an empty array.
+  # The new entry contains only the fields that have non-empty values, to omit optional values that were not provided.
+  ([ -f "${FILENAME}" ] && cat "${FILENAME}" || echo '[]') | \
+      jq \
+        --arg timestamp "$(date +"%Y-%m-%dT%H:%M:%SZ")" \
+        --arg preCommand "${PRE_COMMAND-}" \
+        --arg postCommand "${POST_COMMAND-}" \
+        --arg stage "${STAGE_NAME-}" \
+        --arg phase "${PHASE}" \
+        --arg result "${RESULT-}" \
+        --arg errorLine "${ERROR_LINE-}" \
+        --arg errorMessage "${ERROR_MESSAGE-}" \
+        '. += [
+          {$timestamp,$preCommand,$postCommand,$stage,$phase,$result,$errorLine,$errorMessage} |
+          reduce keys[] as $k (.; if .[$k] == "" then del(.[$k]) else . end)
+        ]' \
+      > "${FILENAME}.tmp" && \
+    mv "${FILENAME}.tmp" "${FILENAME}"
+}
+
+# record_service_start() records the start of a service.
+record_service_start() {
+  if [ "${PRE_COMMAND-}" ]
+  then
+    local PHASE="pre-command start"
+  elif [ "${POST_COMMAND-}" ]
+  then
+    local PHASE="post-command start"
+  else
+    local PHASE="service start"
+  fi
+
+  add_service_record_entry
+}
+
+# record_service_end(result) records the end of a service.
+#   ERROR_LINE - line where the error occurred, if there was an error
+#   ERROR_MESSAGE - error message, if there was an error
+record_service_end() {
+  if [ "${PRE_COMMAND-}" ]
+  then
+    local PHASE="pre-command end"
+  elif [ "${POST_COMMAND-}" ]
+  then
+    local PHASE="post-command end"
+  else
+    local PHASE="service end"
+  fi
+  local RESULT=${1:?Must specify a result}
+
+  add_service_record_entry
+}
+
+# record_service_stage_start(stage_name) records the start of a stage of a service.
+record_service_stage_start() {
+  if [ "${STAGE_NAME-}" ]
+  then
+    echo "attempt to record the start of a stage without ending the previous one"
+    exit 1
+  fi
+
+  local PHASE="stage start"
+  STAGE_NAME=${1:?Must specify a stage name}
+
+  add_service_record_entry
+}
+
+# record_service_stage_end(result) records the end  of a stage of a service.
+#   ERROR_LINE - line where the error occurred, if there was an error
+#   ERROR_MESSAGE - error message, if there was an error
+record_service_stage_end() {
+  if [ -z "${STAGE_NAME-}" ]
+  then
+    echo "attempt to record the end of a stage without starting one"
+    exit 1
+  fi
+
+  local PHASE="stage end"
+  local RESULT=${1:?Must specify a result}
+
+  add_service_record_entry
+
+  STAGE_NAME=""
+}
+
+# record_service_stage_success records the successful end of a stage of a service.
+record_service_stage_success() {
+  record_service_stage_end "success"
+}
+
+record_service_stage_failure() {
+  local ERROR_LINE
+  local ERROR_MESSAGE
+  get_error_info ERROR_LINE ERROR_MESSAGE
+  record_service_stage_end "failure"
+}
+
+record_service_exit() {
+  if [ "$1" -eq 0 ]
+  then
+    local RESULT="success"
+  else
+    local RESULT="failure"
+    local ERROR_LINE
+    local ERROR_MESSAGE
+    get_error_info ERROR_LINE ERROR_MESSAGE
+  fi
+
+  if [ "${STAGE_NAME-}" ]
+  then
+    record_service_stage_end "${RESULT}"
+  fi
+
+  record_service_end "${RESULT}"
+}
+
+get_error_info() {
+  local -n error_line=$1
+  local -n error_message=$2
+  # shellcheck disable=SC2034  # variable used indirectly
+  error_line="$(caller 1)"
+  # shellcheck disable=SC2034  # variable used indirectly
+  error_message="$(journalctl --unit="${SYSTEMD_UNIT_NAME}" --lines=3 --output=cat)"
+}
+
+record_service_start
+
+trap 'record_service_exit $?' EXIT
\ No newline at end of file
diff --git a/data/data/bootstrap/files/usr/local/bin/crio-configure.sh.template b/data/data/bootstrap/files/usr/local/bin/crio-configure.sh.template
index a6c0175f167..28bc1d2d8bd 100755
--- a/data/data/bootstrap/files/usr/local/bin/crio-configure.sh.template
+++ b/data/data/bootstrap/files/usr/local/bin/crio-configure.sh.template
@@ -8,6 +8,8 @@ set -euo pipefail
 # Perhaps down the line we change this to run something like:
 # podman run machine-config-daemon bootstrap ... (passing the release image and the host rootfs)
 
+. /usr/local/bin/bootstrap-service-record.sh
+
 . /usr/local/bin/release-image.sh
 
 MACHINE_CONFIG_INFRA_IMAGE=$(image_for pod)
diff --git a/data/data/bootstrap/files/usr/local/bin/installer-gather.sh b/data/data/bootstrap/files/usr/local/bin/installer-gather.sh
index e6f3802db58..c9f99a0e8cc 100755
--- a/data/data/bootstrap/files/usr/local/bin/installer-gather.sh
+++ b/data/data/bootstrap/files/usr/local/bin/installer-gather.sh
@@ -11,6 +11,10 @@ mkdir -p "${ARTIFACTS}"
 
 exec &> >(tee "${ARTIFACTS}/gather.log")
 
+echo "Gathering bootstrap service records ..."
+mkdir -p "${ARTIFACTS}/bootstrap/services"
+sudo cp -r /var/log/openshift/* "${ARTIFACTS}/bootstrap/services/"
+
 echo "Gathering bootstrap systemd summary ..."
 LANG=POSIX systemctl list-units --state=failed >& "${ARTIFACTS}/failed-units.txt"
 
diff --git a/data/data/bootstrap/files/usr/local/bin/kubelet-pause-image.sh.template b/data/data/bootstrap/files/usr/local/bin/kubelet-pause-image.sh.template
index cffd8ee1aa9..93a318488a8 100755
--- a/data/data/bootstrap/files/usr/local/bin/kubelet-pause-image.sh.template
+++ b/data/data/bootstrap/files/usr/local/bin/kubelet-pause-image.sh.template
@@ -5,6 +5,9 @@ set -euo pipefail
 # Need to set the --pod-infra-container-image flag for the kubelet to point to the pause image from the payload
 # So we add MACHINE_CONFIG_INFRA_IMAGE to an environment file and source that in the kubelet service
 
+PRE_COMMAND="kubelet-pause-image"
+. /usr/local/bin/bootstrap-service-record.sh
+
 . /usr/local/bin/release-image.sh
 
 echo "MACHINE_CONFIG_INFRA_IMAGE=$(image_for pod)" > /etc/kubernetes/kubelet-pause-image-override
diff --git a/data/data/bootstrap/files/usr/local/bin/kubelet.sh b/data/data/bootstrap/files/usr/local/bin/kubelet.sh
new file mode 100644
index 00000000000..9af2b827202
--- /dev/null
+++ b/data/data/bootstrap/files/usr/local/bin/kubelet.sh
@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+
+# shellcheck disable=SC1091  # using path on bootstrap machine
+. /usr/local/bin/bootstrap-service-record.sh
+
+/usr/bin/hyperkube \
+  kubelet \
+    --anonymous-auth=false \
+    --container-runtime=remote \
+    --container-runtime-endpoint=/var/run/crio/crio.sock \
+    --runtime-request-timeout="${KUBELET_RUNTIME_REQUEST_TIMEOUT}" \
+    --pod-manifest-path=/etc/kubernetes/manifests \
+    --minimum-container-ttl-duration=6m0s \
+    --cluster-domain=cluster.local \
+    --cgroup-driver=systemd \
+    --serialize-image-pulls=false \
+    --v=2 \
+    --volume-plugin-dir=/etc/kubernetes/kubelet-plugins/volume/exec \
+    --pod-infra-container-image="${MACHINE_CONFIG_INFRA_IMAGE}"
diff --git a/data/data/bootstrap/files/usr/local/bin/release-image-download.sh.template b/data/data/bootstrap/files/usr/local/bin/release-image-download.sh.template
index 71a73484d29..d96c793faf2 100755
--- a/data/data/bootstrap/files/usr/local/bin/release-image-download.sh.template
+++ b/data/data/bootstrap/files/usr/local/bin/release-image-download.sh.template
@@ -9,11 +9,20 @@ set -euo pipefail
 # service: https://github.com/systemd/systemd/issues/2582.
 #
 
+. /usr/local/bin/bootstrap-service-record.sh
+
 RELEASE_IMAGE={{.ReleaseImage}}
 
 echo "Pulling $RELEASE_IMAGE..."
-while ! podman pull --quiet "$RELEASE_IMAGE"
+while true
 do
-    echo "Pull failed. Retrying $RELEASE_IMAGE..."
+    record_service_stage_start "pull-release-image"
+    if podman pull --quiet "$RELEASE_IMAGE"
+    then
+        record_service_stage_success
+        break
+    else
+        record_service_stage_failure
+        echo "Pull failed. Retrying $RELEASE_IMAGE..."
+    fi
 done
-
diff --git a/data/data/bootstrap/files/usr/local/bin/report-progress.sh b/data/data/bootstrap/files/usr/local/bin/report-progress.sh
index 16f662c63a0..7f625440aca 100755
--- a/data/data/bootstrap/files/usr/local/bin/report-progress.sh
+++ b/data/data/bootstrap/files/usr/local/bin/report-progress.sh
@@ -1,17 +1,23 @@
 #!/usr/bin/env bash
 
+# shellcheck disable=SC1091  # using path on bootstrap machine
+. /usr/local/bin/bootstrap-service-record.sh
+
 KUBECONFIG="${1}"
 
-wait_for_existance() {
+wait_for_existence() {
 	while [ ! -e "${1}" ]
 	do
 		sleep 5
 	done
 }
 
+record_service_stage_start "wait-for-bootstrap-complete"
 echo "Waiting for bootstrap to complete..."
-wait_for_existance /opt/openshift/.bootkube.done
+wait_for_existence /opt/openshift/.bootkube.done
+record_service_stage_success
 
+record_service_stage_start "report-bootstrap-complete"
 echo "Reporting install progress..."
 while ! oc --kubeconfig="$KUBECONFIG" create -f - <<-EOF
 	apiVersion: v1
@@ -25,3 +31,4 @@ EOF
 do
 	sleep 5
 done
+record_service_stage_success
diff --git a/data/data/bootstrap/systemd/units/chown-gatewayd-key.service b/data/data/bootstrap/systemd/units/chown-gatewayd-key.service
index 993414fbc72..d226bc4a2d0 100644
--- a/data/data/bootstrap/systemd/units/chown-gatewayd-key.service
+++ b/data/data/bootstrap/systemd/units/chown-gatewayd-key.service
@@ -9,7 +9,7 @@ Before=systemd-journal-gatewayd.service
 [Service]
 Type=oneshot
 RemainAfterExit=yes
-ExecStart=/bin/sh -c "if ! getent passwd systemd-journal-gateway &>/dev/null; then useradd -r systemd-journal-gateway; fi && chown systemd-journal-gateway: /opt/openshift/tls/journal-gatewayd.{crt,key}"
+ExecStart=/bin/sh -c ". /usr/local/bin/bootstrap-service-record.sh; if ! getent passwd systemd-journal-gateway &>/dev/null; then useradd -r systemd-journal-gateway; fi && chown systemd-journal-gateway: /opt/openshift/tls/journal-gatewayd.{crt,key}"
 
 [Install]
 WantedBy=multi-user.target
diff --git a/data/data/bootstrap/systemd/units/kubelet.service.template b/data/data/bootstrap/systemd/units/kubelet.service.template
index 1888ddeea8e..092d4c8e6e1 100644
--- a/data/data/bootstrap/systemd/units/kubelet.service.template
+++ b/data/data/bootstrap/systemd/units/kubelet.service.template
@@ -5,6 +5,7 @@ After=crio.service release-image.service
 
 [Service]
 Type=notify
+NotifyAccess=all
 ExecStartPre=/bin/mkdir --parents /etc/kubernetes/manifests
 ExecStartPre=/bin/mkdir --parents /etc/kubernetes/kubelet-plugins/volume/exec
 ExecStartPre=/usr/local/bin/kubelet-pause-image.sh
@@ -12,20 +13,7 @@ Environment=KUBELET_RUNTIME_REQUEST_TIMEOUT=10m
 EnvironmentFile=-/etc/kubernetes/kubelet-env
 EnvironmentFile=-/etc/kubernetes/kubelet-pause-image-override
 
-ExecStart=/usr/bin/hyperkube \
-  kubelet \
-    --anonymous-auth=false \
-    --container-runtime=remote \
-    --container-runtime-endpoint=/var/run/crio/crio.sock \
-    --runtime-request-timeout=${KUBELET_RUNTIME_REQUEST_TIMEOUT} \
-    --pod-manifest-path=/etc/kubernetes/manifests \
-    --minimum-container-ttl-duration=6m0s \
-    --cluster-domain=cluster.local \
-    --cgroup-driver=systemd \
-    --serialize-image-pulls=false \
-    --v=2 \
-    --volume-plugin-dir=/etc/kubernetes/kubelet-plugins/volume/exec \
-    --pod-infra-container-image=${MACHINE_CONFIG_INFRA_IMAGE}
+ExecStart=/usr/local/bin/kubelet.sh
 
 Restart=always
 RestartSec=10
diff --git a/docs/dev/bootstrap_services.md b/docs/dev/bootstrap_services.md
new file mode 100644
index 00000000000..97759a9e808
--- /dev/null
+++ b/docs/dev/bootstrap_services.md
@@ -0,0 +1,95 @@
+#### Bootstrap Service Records ####
+
+For the purposes of diagnosing installation failures that occur during bootstrapping, the
+progresses of services running on the bootstrap machines are tracked in json files in the
+/var/log/openshift directory. The progress for each service is tracked in its own file. For
+example, the bootkube service progress is tracked in the /var/log/openshift/bootkube.json
+file. The following progress events are tracked.
+* A service adds an entry when the service starts.
+* A service adds an entry when the service ends. The entry includes the result of the service
+invocation, either success or failure. If the invocation failed, then the entry includes
+the line number of the error and the last three lines from the service's journal log.
+* A service adds an entry when a stage of the service starts. An example of a service stage
+is the cvo-bootstrap stage of the bootkube service. During that stage, the
+cluster-version-operator renders its manifests.
+* A service adds an entry when a stage of the service ends. This is similar to the entry
+added when a service ends, including having a result and error information if applicable.
+* A service adds an entry when a pre- or post-command starts.
+* A service adds en entry when a pre- or post-command ends. This is similar to the entry
+added when a service ends, including having a result and error information if applicable.
+
+##### Managing Service Records in a Service #####
+
+To track its progress, a service should source the /usr/local/bin/bootstrap-service-record.sh
+script. When a service sources the script, the script will add an entry to the json file for
+the service indicating that the service started. When the service ends, either successfully
+or due to an error, the script will add an entry to the json file for the service indicating that
+the service ended. The script will consider whether the last command executed was successful or
+not in order to determine whether the service was successful.
+
+For tracking stages, the service should call functions from the sourced script.
+* The service should call the `record_service_stage_start` function when a stage of the service
+starts. The function takes the name of the stage as its single argument.
+* The service should call the `record_service_stage_success` function when a stage of the service
+ends successfully.
+* The service should call the `record_service_stage_failure` function when a stage of the service
+ends due to a failure. The script will automatically record an entry for a stage failure if the
+service ends during the execution of a stage.
+
+###### Pre- and Post-Commands ######
+
+If a service has pre- or post-commands that could either run for significant periods or could
+potentially fail, then those commands should add to the json file as well. Such a command should
+source the same /usr/local/bin/bootstrap-service-record.sh script. It should also set either the
+`PRE_COMMAND` or `POST_COMMAND` environment variable with a value that identifies the command.
+For example, kubelet.service has a pre-command of /usr/local/bin/kubelet-pause-image.sh. The
+kubelet-pause-image.sh script sets the `PRE_COMMAND` environment variable to "kubelet-pause-image"
+before sourcing the bootstrap-service-record.sh script. The entries for the pre-command will
+contain a `preCommand` field with the "kubelet-pause-image" value.
+
+###### Sample Script #######
+
+```shell script
+#!/usr/bin/env bash
+set -euoE pipefail
+
+# Source the script to record service entries.
+# This will create en entry for the start of the service.
+. /usr/local/bin/bootstrap-service-record.sh
+
+# Record the start of the "first" stage.
+record_service_stage_start "first"
+
+# Record the successful end of the "first" stage.
+record_service_stage_success
+
+while true
+do
+    # Record the start of the "second" stage.
+    record_service_stage_start "second-stage"
+    
+    if [ some_check ]
+    then
+        # Record the successful end of the "second" stage.
+        record_service_stage_success
+        break
+    else
+        # Record the failing end of the "second" stage.
+        record_service_stage_failure  
+    fi
+done
+
+# Record the start of the third stage.
+record_service_stage_start "third"
+
+# If the command fails, then an entry will be recorded for the failing end
+# of the third stage and an entry will be recorded for the failing end of
+# the service.
+some_command_that_may_fail
+
+# Record the end of the third stage.
+record_service_stage_success
+
+# Since this is the end of the script, an entry will be recorded for the
+# successful end of the service.
+```
\ No newline at end of file