Add Scrutiny Helm chart by votdev · Pull Request #502 · harvester/charts

votdev · 2026-03-26T11:12:56Z

Problem:

Prometheus node-exporter does not collect SMART values of the disks connected to the cluster nodes.

Solution:

Make use of the scrutiny project which provides a nice small UI to display the SMART values of the disks that are auto-detected on each node in the cluster.

Related Issue(s):

n/a

Test plan:

Download the Helm chart

curl -sL "https://github.com/votdev/harvester-charts/archive/refs/heads/scrutiny.tar.gz" \
  | tar -xz --strip-components=2 harvester-charts-scrutiny/charts/scrutiny

Validate the Helm chart

helm template scrutiny ./scrutiny -n scrutiny
helm lint ./scrutiny -f values.yaml
helm install scrutiny ./scrutiny -n scrutiny --dry-run --debug

Install the Helm chart

helm install scrutiny -n harvester-system --dry-run ./scrutiny
helm install scrutiny -n harvester-system ./scrutiny

Validate installation (on a 3 node cluster)
Check the running pods.
3 collector pods
1 Influx DB pod
1 Web pod

kubectl get deployment -n harvester-system

Go to Service Discovery > Services in the embedded Rancher. Look out for scrutiny-web. Click on the IP in the Targetcolumn to open the Scrutiny WebUI.

Additional documentation or context

https://suse.slack.com/archives/C02D9459KFW/p1773321565394789

Copilot

Pull request overview

Adds a new scrutiny Helm chart to deploy the Scrutiny stack (web UI, collector DaemonSet, and InfluxDB) so SMART disk metrics can be collected and viewed in-cluster.

Changes:

Introduces a new Helm chart (charts/scrutiny) with default values and chart metadata.
Adds Kubernetes manifests for Scrutiny Web + InfluxDB Deployments, Services, PVCs, and a collector DaemonSet.
Adds Helm helper templates for naming/namespace overrides.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 13 comments.

Show a summary per file

File	Description
charts/scrutiny/Chart.yaml	Declares the new Helm chart metadata (name/version/appVersion).
charts/scrutiny/values.yaml	Provides default images, persistence sizes, tolerations, and collector scheduling.
charts/scrutiny/templates/_helpers.tpl	Adds helper templates for name/fullname/namespace.
charts/scrutiny/templates/deployment-web.yaml	Deploys the Scrutiny web UI and config PVC mount.
charts/scrutiny/templates/service-web.yaml	Exposes the web UI via a ClusterIP Service.
charts/scrutiny/templates/deployment-influxdb.yaml	Deploys InfluxDB with a data PVC mount.
charts/scrutiny/templates/service-influxdb.yaml	Exposes InfluxDB internally to the web/collector components.
charts/scrutiny/templates/pvc-influxdb.yaml	Creates persistent storage for InfluxDB data.
charts/scrutiny/templates/pvc-config.yaml	Creates persistent storage for Scrutiny config.
charts/scrutiny/templates/daemonset-collector.yaml	Runs the Scrutiny collector on each node with host access for SMART collection.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 8 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Signed-off-by: Volker Theile <vtheile@suse.com>

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-26T13:14:51Z

+              mountPath: /opt/scrutiny/config
+      volumes:
+        - name: config
+          persistentVolumeClaim:
+            claimName: {{ include "scrutiny.fullname" . }}-config-pvc


The web Deployment mounts a single ReadWriteOnce PVC (config-pvc) but allows scaling via replicas. Scaling above 1 will leave extra replicas Pending due to the RWO volume, and the default RollingUpdate strategy can also stall rollouts when a surge pod tries to mount the same PVC. Consider constraining this component to 1 replica and/or setting spec.strategy.type: Recreate (or switching to StatefulSet / RWX storage) to make updates reliable.

Copilot · 2026-03-26T13:14:52Z

+              mountPath: /var/lib/influxdb2
+      volumes:
+        - name: influxdb-data
+          persistentVolumeClaim:
+            claimName: {{ include "scrutiny.fullname" . }}-influxdb-pvc


This Deployment mounts a single ReadWriteOnce PVC for /var/lib/influxdb2 but does not set spec.strategy. With the default RollingUpdate strategy, upgrades can stall because a surge pod may be created before the old pod is terminated, and both cannot mount the same RWO volume. Set spec.strategy.type: Recreate (see charts/harvester-vm-import-controller/templates/deployment.yaml:9-10) or use a StatefulSet to make rollouts reliable.

Copilot · 2026-03-26T13:14:52Z

+  - effect: NoSchedule
+    key: node-role.kubernetes.io/master
+  - effect: NoExecute
+    operator: Exists


The default toleration effect: NoExecute with operator: Exists and no tolerationSeconds tolerates all NoExecute taints indefinitely (including node.kubernetes.io/not-ready / unreachable). For the web/influxdb Deployments this can reduce availability by preventing eviction/rescheduling when a node becomes unhealthy. Consider scoping this toleration to the specific taints you need, or adding tolerationSeconds for node-condition taints.

Suggested change

operator: Exists

operator: Exists

tolerationSeconds: 300

w13915984028 · 2026-03-27T07:54:29Z

Thanks for the PR, some open questions for discussion,

(1) How to represent this chart to user.

It could be a subchart of rancher-monitoring https://github.com/rancher/charts/tree/dev-v2.14/charts/rancher-monitoring ; or an addon/experimental-addon on Harvester.

Rancher mirror all/most of the images by third-party like https://github.com/harvester/harvester-installer/blob/c8fd2f4e80cdeef1d79396ef75c6e49deb4d0ba8/scripts/images/rancher-images.txt#L20

(2) License https://github.com/analogj/scrutiny?tab=readme-ov-file#licenses, I am not an expert on this area.

(3) Test plans, coutinuous upgrade plans.

maybe we need a HEP.

innobead · 2026-05-27T08:04:55Z

As discussed, we should consider enhancing our node disk manager to provide enough capacity to monitor disk statistics and export them via our own Prometheus metrics API. When using an upstream chart, consider maintenance, resource consumption, release workload, and feature scope, among other factors. This requires strong justification and analysis to ensure alignment with our project goals. (Otherwise, users can simply install the upstream chart on their own, since there are no integration benefits here.)

Suggest reviewing harvester/harvester#7416 (comment) first for ideas from Longhorn SMART monitoring (a minimal implementation of the required capabilities), including kernel-managed disks, filesystems, or user-space disks managed by SPDK, to help shape our design. Once you review this, you'll have a clearer understanding of the effort involved and whether any open-source projects could help us.

cc @harvester/storage @harvester/pm

votdev self-assigned this Mar 26, 2026

Copilot AI review requested due to automatic review settings March 26, 2026 11:12

Copilot started reviewing on behalf of votdev March 26, 2026 11:13 View session

votdev force-pushed the scrutiny branch from 57bcc11 to 4cc668e Compare March 26, 2026 11:14

Copilot AI reviewed Mar 26, 2026

View reviewed changes

votdev requested review from Vicente-Cheng, WebberHuang1118, ibrokethecloud, irishgordo and w13915984028 March 26, 2026 11:38

votdev force-pushed the scrutiny branch from 4cc668e to b16c3bb Compare March 26, 2026 11:59

votdev requested a review from Copilot March 26, 2026 12:02

Copilot started reviewing on behalf of votdev March 26, 2026 12:02 View session

votdev force-pushed the scrutiny branch from b16c3bb to c24cdd7 Compare March 26, 2026 12:07

Copilot AI reviewed Mar 26, 2026

View reviewed changes

Comment thread charts/scrutiny/templates/pvc-influxdb.yaml

Comment thread charts/scrutiny/templates/pvc-config.yaml

votdev requested a review from Copilot March 26, 2026 12:11

Copilot started reviewing on behalf of votdev March 26, 2026 12:11 View session

Copilot AI reviewed Mar 26, 2026

View reviewed changes

Add scrutiny Helm chart

2665ce5

Signed-off-by: Volker Theile <vtheile@suse.com>

votdev force-pushed the scrutiny branch from c24cdd7 to 2665ce5 Compare March 26, 2026 12:54

votdev requested a review from Copilot March 26, 2026 13:07

Copilot started reviewing on behalf of votdev March 26, 2026 13:08 View session

Copilot AI reviewed Mar 26, 2026

View reviewed changes

Uh oh!

Conversation

votdev commented Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem:

Solution:

Related Issue(s):

Test plan:

Additional documentation or context

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

w13915984028 commented Mar 27, 2026

Uh oh!

innobead commented May 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

votdev commented Mar 26, 2026 •

edited

Loading