From 8e45a101a1a5355d2ebc1be63505eff3cf9566a5 Mon Sep 17 00:00:00 2001
From: paulfantom <pawel@krupa.net.pl>
Date: Fri, 22 Nov 2024 13:43:06 +0100
Subject: [PATCH] npd: test a solution to a bug in npd

---
 base/node-problem-detector/repository.yaml |  3 +-
 base/node-problem-detector/values.yaml     | 42 +++++++++++++++++++++-
 2 files changed, 43 insertions(+), 2 deletions(-)

diff --git a/base/node-problem-detector/repository.yaml b/base/node-problem-detector/repository.yaml
index 967b38f49..8777a12af 100644
--- a/base/node-problem-detector/repository.yaml
+++ b/base/node-problem-detector/repository.yaml
@@ -5,4 +5,5 @@ metadata:
   namespace: node-problem-detector
 spec:
   interval: 5m
-  url: https://charts.deliveryhero.io/
+  oci: true
+  url: oci://ghcr.io/deliveryhero/helm-charts
diff --git a/base/node-problem-detector/values.yaml b/base/node-problem-detector/values.yaml
index b1bc5dede..250580956 100644
--- a/base/node-problem-detector/values.yaml
+++ b/base/node-problem-detector/values.yaml
@@ -20,7 +20,47 @@ resources:
     memory: 18Mi
 
 settings:
+  #log_monitors:
+  #  - /config/kernel-monitor.json
+  #  - /config/readonly-monitor.json
+  #  - /config/health-checker-containerd.json
   log_monitors:
     - /config/kernel-monitor.json
     - /config/readonly-monitor.json
-    - /config/health-checker-containerd.json
+    - /custom-config/docker-monitor-filelog.json
+
+  custom_monitor_definitions:
+    health-checker-containerd.json: |
+      {
+        "plugin": "custom",
+        "pluginConfig": {
+          "invoke_interval": "10s",
+          "timeout": "3m",
+          "max_output_length": "80",
+          "concurrency": "1"
+        },
+        "source": "health-checker",
+        "metricsReporting": true,
+        "conditions": [
+          {
+            "type": "ContainerRuntimeUnhealthy",
+            "reason": "ContainerRuntimeIsHealthy",
+            "message": "Container runtime on the node is functioning properly"
+          }
+        ],
+        "rules": [
+          {
+            "type": "permanent",
+            "condition": "ContainerRuntimeUnhealthy",
+            "reason": "ContainerdUnhealthy",
+            "path": "/home/kubernetes/bin/health-checker",
+            "args": [
+              "--component=cri",
+              "--enable-repair=true",
+              "--cooldown-time=2m",
+              "--health-check-timeout=60s"
+            ],
+            "timeout": "3m"
+          }
+        ]
+      }