Implement Health checks (#178)

* add healthchecks to moonraker, klipper and ustreamer images

CHANGELOG.md

@@ -8,6 +8,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 ## [Unreleased]
 ### Added
 * klipper & moonraker: generate version file during build to correctly display versions
+* klipper, moonraker & ustreamer: add healthchecks to container images
 ### Fixed
 ### Changed
 ### Removed

docker/klipper/Dockerfile

@@ -39,6 +39,9 @@ RUN groupadd klipper --gid 1000 \
 RUN mkdir -p printer_data/run printer_data/gcodes printer_data/logs printer_data/config \
  && chown -R klipper:klipper /opt/*
 
+COPY --chown=klipper:klipper health.py ./
+HEALTHCHECK --interval=5s CMD ["python3", "/opt/health.py"]
+
 COPY --chown=klipper:klipper --from=build /opt/klipper ./klipper
 COPY --chown=klipper:klipper --from=build /opt/venv ./venv
 

docker/klipper/README.md

@@ -118,4 +118,11 @@ none
 |`build-hostmcu`|Based on `mcu`: Build the klipper_mcu binary|No|
 |`run`|Default runtime Image for klippy|Yes|
 |`tools`|Build Tools for MCU code compilation|Yes|
-|`hostmcu`|Runtime Image for the klipper_mcu binary|Yes|
+|`hostmcu`|Runtime Image for the klipper_mcu binary|Yes|
+
+## Healthcheck
+`/opt/health.py` gets executed every 5s inside the container.  
+The script does the following:
+* queries klippers `info` endpoint via its unix socket
+* Checks if state is `ready`
+* If one of the above requirements is not `ready`, the script exits with a failure state to indicate the container is unhealthy

docker/klipper/health.py

@@ -0,0 +1,22 @@
+#!/usr/bin/env python3
+import socket, json, sys
+
+socket_address="/opt/printer_data/run/klipper.sock"
+message={"id": 666, "method": "info"}
+
+# Set up socket connection
+sock = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM)
+sock.connect(socket_address)
+
+# Send message and receive response
+sock.sendall(json.dumps(message).encode() + b"\x03")
+response = sock.recv(4096).decode('utf-8').strip('\x03')
+sock.close()
+
+# Check the result
+if json.loads(response)["result"]["state"] == "ready":
+    # State is ready - healthy
+    sys.exit(0)
+else:
+    # State is not ready - unhealthy
+    sys.exit(1)

docker/moonraker/Dockerfile

@@ -39,6 +39,7 @@ RUN apt update \
       systemd \
       sudo \
       git \
+      jq \
  && apt clean
 
 WORKDIR /opt
@@ -48,6 +49,9 @@ RUN groupadd moonraker --gid 1000 \
 RUN mkdir -p printer_data/run printer_data/gcodes printer_data/logs printer_data/database printer_data/config \
  && chown -R moonraker:moonraker /opt/*
 
+COPY --chown=moonraker:moonraker health.sh ./
+HEALTHCHECK --interval=5s CMD ["bash", "/opt/health.sh"]
+
 COPY --chown=moonraker:moonraker --from=build /opt/moonraker ./moonraker
 COPY --chown=moonraker:moonraker --from=build /opt/venv ./venv
 

docker/moonraker/README.md

@@ -84,4 +84,14 @@ services:
 |Target|Description|Pushed|
 |---|---|---|
 |`build`|Pull Upstream Codebase and build python venv|No|
-|`run`|Default runtime Image|Yes|
+|`run`|Default runtime Image|Yes|
+
+## Healthcheck
+`/opt/health.sh` gets executed every 5s inside the container.  
+The script does the following:
+* queries the `/server/info` endpoint of moonraker
+* Performs the following checks
+  * Number of failed moonraker_components = 0
+  * klippy_connected is `true`
+  * klippy_state is `ready`
+* If one of the above requirements is not met, the script exits with a failure state to indicate the container is unhealthy

docker/moonraker/health.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+serverinfo=$(curl -s localhost:7125/server/info)
+
+klippy_connected=$(echo -n ${serverinfo} | jq -r .result.klippy_connected)
+klippy_state=$(echo -n ${serverinfo} | jq -r .result.klippy_state)
+failed_components=$(echo -n ${serverinfo} | jq -r .result.failed_components[] | wc -l)
+
+if [ "$klippy_connected" == "true" ] \
+&& [ "$klippy_state" == "ready" ] \
+&& [ $failed_components -eq 0 ]; then
+  ## moonraker is up and connected to klippy
+  exit 0
+else
+  ## moonraker started w/ failed components and/or is not connected to klippy
+  exit 1
+fi

docker/ustreamer/Dockerfile

@@ -39,6 +39,8 @@ RUN apt update \
       libbsd0 \
       libgpiod2 \
       v4l-utils \
+      curl \
+      jq \
  && apt clean
 
 WORKDIR /opt
@@ -46,11 +48,14 @@ RUN groupadd ustreamer --gid 1000 \
  && useradd ustreamer --uid 1000 --gid ustreamer \
  && usermod ustreamer --append --groups video
 
+COPY --chown=ustreamer:ustreamer health.sh ./
+HEALTHCHECK --interval=5s CMD ["bash", "/opt/health.sh"]
+
 COPY --chown=ustreamer:ustreamer --from=build /opt/ustreamer/src/ustreamer.bin ./ustreamer
 
 ## Start ustreamer
 USER ustreamer
 EXPOSE 8080
-ENTRYPOINT [ "/opt/ustreamer"]
+ENTRYPOINT ["/opt/ustreamer"]
 CMD ["--host=0.0.0.0", "--port=8080"]
 

docker/ustreamer/README.md

@@ -51,4 +51,13 @@ none
 |Target|Description|Pushed|
 |---|---|---|
 |`build`|Pull Upstream Codebase and build application|No|
-|`run`|Default runtime Image|Yes|
+|`run`|Default runtime Image|Yes|
+
+## Healthcheck
+`/opt/health.sh` gets executed every 5s inside the container.  
+The script does the following:
+* gets the JSON structure with the state of the server
+* Checks the following values
+  * `.ok` is set to `true`, which indicates ustreamer is working
+  * `.result.source.online` is set to `true`, which indicates the source (webcam) is returning an image rather than `NO SIGNAL`
+* If one of the above requirements is not met, the script exits with a failure state to indicate the container is unhealthy

docker/ustreamer/health.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+state=$(curl -s localhost:8080/state)
+ok=$(echo $state | jq -r .ok)
+online=$(echo $state | jq -r .result.source.online)
+
+if [ "$ok" == "true" ] && [ "$online" == "true" ]; then
+  ## ustreamer is ok and source is online
+  exit 0
+else
+  ## ustreamer is not ok or source is not online
+  exit 1
+fi