Merge pull request #66 from floushee/main

Added support for Slack / MS Teams webhooks

Dockerfile

@@ -10,7 +10,8 @@ ENV AUTOHEAL_CONTAINER_LABEL=autoheal \
     AUTOHEAL_INTERVAL=5 \
     AUTOHEAL_DEFAULT_STOP_TIMEOUT=10 \
     DOCKER_SOCK=/var/run/docker.sock \
-    CURL_TIMEOUT=30
+    CURL_TIMEOUT=30 \
+    WEBHOOK_URL=""
 
 HEALTHCHECK --interval=5s CMD pgrep -f autoheal || exit 1
 

README.md

@@ -60,6 +60,7 @@ AUTOHEAL_START_PERIOD=0   # wait 0 seconds before first health check
 AUTOHEAL_DEFAULT_STOP_TIMEOUT=10   # Docker waits max 10 seconds (the Docker default) for a container to stop before killing during restarts (container overridable via label, see below)
 DOCKER_SOCK=/var/run/docker.sock   # Unix socket for curl requests to Docker API
 CURL_TIMEOUT=30     # --max-time seconds for curl requests to Docker API
+WEBHOOK_URL=""    # post message to the webhook if a container was restarted (or restart failed)
 ```
 
 ### Optional Container Labels

docker-entrypoint

@@ -7,6 +7,7 @@ set -o pipefail
 DOCKER_SOCK=${DOCKER_SOCK:-/var/run/docker.sock}
 UNIX_SOCK=""
 CURL_TIMEOUT=${CURL_TIMEOUT:-30}
+WEBHOOK_URL=${WEBHOOK_URL:-""}
 
 # only use unix domain socket if no TCP endpoint is defined
 case "${DOCKER_SOCK}" in
@@ -56,6 +57,26 @@ restart_container() {
   docker_curl -f -X POST "${HTTP_ENDPOINT}/containers/${container_id}/restart?t=${timeout}"
 }
 
+notify_webhook() {
+  local text="$@"
+
+  if [ -n "$WEBHOOK_URL" ]
+  then
+    # execute webhook requests as background process to prevent healer from blocking
+    curl -X POST -H "Content-type: application/json" -d "$(generate_webhook_payload $text)"  $WEBHOOK_URL
+  fi
+}
+
+# https://towardsdatascience.com/proper-ways-to-pass-environment-variables-in-json-for-curl-post-f797d2698bf3
+generate_webhook_payload() {
+  local text="$@"
+  cat <<EOF
+{
+  "text":"$text"
+} 
+EOF
+}
+
 # SIGTERM-handler
 term_handler() {
   exit 143  # 128 + 15 -- SIGTERM
@@ -94,6 +115,9 @@ if [ "$1" = "autoheal" ] && [ -e "$DOCKER_SOCK" ];then
         if ! restart_container "$CONTAINER_ID" "$TIMEOUT"
         then
           echo "$DATE Restarting container $CONTAINER_SHORT_ID failed" >&2
+          notify_webhook "Container ${CONTAINER_NAME:1} (${CONTAINER_SHORT_ID}) found to be unhealthy. Failed to restart the container!" &
+        else 
+          notify_webhook "Container ${CONTAINER_NAME:1} (${CONTAINER_SHORT_ID}) found to be unhealthy. Successfully restarted the container!" &
         fi
       fi
     done