From 126af5044b45fd078fc8a413bdb1db58690f3219 Mon Sep 17 00:00:00 2001
From: Solomon Jacobs <solomon.jacobs@checkmk.com>
Date: Wed, 3 Apr 2024 15:17:35 +0200
Subject: [PATCH] 16417 FIX Add Configuration Option 'checkmkAgentTimeout'

CMK-16676

Closes: #26

Change-Id: I697df13efc1d6b5279396d626b0c335c51928892
---
 .werks/16417                                  | 22 +++++++++++++++++++
 .../node-collector-machine-sections-ds.yaml   |  3 +++
 deploy/charts/checkmk/values.yaml             |  2 ++
 src/checkmk_kube_agent/send_metrics.py        |  6 ++---
 4 files changed, 30 insertions(+), 3 deletions(-)
 create mode 100644 .werks/16417

diff --git a/.werks/16417 b/.werks/16417
new file mode 100644
index 0000000..3eb8e8b
--- /dev/null
+++ b/.werks/16417
@@ -0,0 +1,22 @@
+Title: Add Configuration Option 'checkmkAgentTimeout'
+Class: fix
+Compatible: compat
+Component: node-collector
+Date: 1712152033
+Knowledge: doc
+Level: 1
+Version: 2.0.0-alpha.1
+
+The machine-sections-collector executes a version of the 'check_mk_agent' to collect information
+about the host. Sometimes this script takes more than five seconds, which causes the following
+traceback.
+
+C+:
+ File "/usr/local/lib/python3.10/subprocess.py", line 1935, in _wait
+    raise TimeoutExpired(self.args, timeout)
+subprocess.TimeoutExpired: Command '['/usr/local/bin/check_mk_agent']' timed out after 5 seconds
+C-:
+
+If you encounter this error, you can configure a longer timeout via the new option
+'nodeCollector.machineSectionsCollector.checkmkAgentTimeout' in the 'values.yaml' configuration
+file.
diff --git a/deploy/charts/checkmk/templates/node-collector-machine-sections-ds.yaml b/deploy/charts/checkmk/templates/node-collector-machine-sections-ds.yaml
index ac9c89f..31c6471 100644
--- a/deploy/charts/checkmk/templates/node-collector-machine-sections-ds.yaml
+++ b/deploy/charts/checkmk/templates/node-collector-machine-sections-ds.yaml
@@ -61,6 +61,9 @@ spec:
             - "/usr/local/bin/checkmk-machine-sections-collector"
           args:
             - "--log-level={{ .Values.nodeCollector.logLevel }}"
+          {{- if .Values.nodeCollector.machineSectionsCollector.checkmkAgentTimeout }}
+            - "--checkmk-agent-timeout={{ .Values.nodeCollector.machineSectionsCollector.checkmkAgentTimeout }}"
+          {{- end }}
           {{- if .Values.tlsCommunication.enabled }}
             - "--secure-protocol"
           {{- if .Values.tlsCommunication.verifySsl }}
diff --git a/deploy/charts/checkmk/values.yaml b/deploy/charts/checkmk/values.yaml
index 9716e52..b782155 100644
--- a/deploy/charts/checkmk/values.yaml
+++ b/deploy/charts/checkmk/values.yaml
@@ -282,6 +282,8 @@ nodeCollector:
         cpu: 150m
         memory: 200Mi
 
+    checkmkAgentTimeout: 5
+
     # the machine sections collector can collect monitoring information for network interfaces of the underlying node.
     # this means that the '/sys' directory of the node will be mounted into the container.
     # the pod security policy is adjusted accordingly.
diff --git a/src/checkmk_kube_agent/send_metrics.py b/src/checkmk_kube_agent/send_metrics.py
index 2715f27..1a5531d 100644
--- a/src/checkmk_kube_agent/send_metrics.py
+++ b/src/checkmk_kube_agent/send_metrics.py
@@ -303,7 +303,7 @@ def parse_arguments(argv: Sequence[str]) -> argparse.Namespace:
         help="Collector log level.",
     )
     parser.add_argument(
-        "--agent-timeout",
+        "--checkmk-agent-timeout",
         type=int,
         help="Checkmk Agent execution timeout in seconds",
     )
@@ -313,7 +313,7 @@ def parse_arguments(argv: Sequence[str]) -> argparse.Namespace:
         max_retries=10,
         polling_interval=60,
         ca_cert="/etc/ca-certificates/checkmk-ca-cert.pem",
-        agent_timeout=5,
+        checkmk_agent_timeout=5,
     )
 
     return parser.parse_args(argv)
@@ -324,7 +324,7 @@ def container_metrics_worker(
     cluster_collector_base_url: Url,
     headers: RequestHeaders,
     verify: SslVerify,
-    args: argparse.Namespace,  # pylint: disable=unused-argument
+    _args: argparse.Namespace,
 ) -> None:  # pragma: no cover
     """
     Query cadvisor api, send metrics to cluster collector