Implement PriorityClass Support

[kaovilai]

Add priority class support for Velero components

Summary

This PR implements priority class support for Velero components, allowing users to configure Kubernetes priority classes for the Velero server, node-agent, data mover pods, and maintenance jobs. This ensures proper pod scheduling and preemption behavior in resource-constrained environments.

Changes

Core Implementation

• Added --server-priority-class-name and --node-agent-priority-class-name flags to velero install command
• Modified node-agent server to read priority class from ConfigMap once at startup (minimizing API calls)
• Updated all data mover controllers to accept priority class as a string parameter
• Added priority class validation with fallback behavior to prevent pod creation failures
• Removed ConfigProvider infrastructure in favor of static configuration reading

Key Features

• Validation with Fallback: ValidatePriorityClass now returns a boolean, allowing callers to handle missing priority classes gracefully
• Single ConfigMap Read: Node-agent reads ConfigMap once at startup and passes configuration to controllers
• Automatic Fallback: If a priority class doesn't exist, the system logs a warning and uses default scheduling

Configuration Flow

graph TB
    subgraph "Installation Time"
        CLI[velero install] --> |"--server-priority-class-name"| Server[Velero Server Deployment]
        CLI --> |"--node-agent-priority-class-name"| NodeAgent[Node Agent DaemonSet]
    end
    
    subgraph "ConfigMap Configuration"
        NAConfig[node-agent-configmap] --> |priorityClassName| DataMovers[Data Mover Pods]
        MJConfig[repo-maintenance-job-configmap] --> |global.priorityClassName| MaintJobs[Maintenance Jobs]
    end
    
    subgraph "Runtime Flow"
        NodeAgentServer[Node Agent Server Startup] --> |Read Once| NAConfig
        NodeAgentServer --> |Validate| ValidatePC{Priority Class Exists?}
        ValidatePC -->|Yes| PassToControllers[Pass to Controllers]
        ValidatePC -->|No| ClearPC[Clear Priority Class<br/>Log Warning]
        ClearPC --> PassToControllers
        PassToControllers --> PVBController[PVB Controller]
        PassToControllers --> PVRController[PVR Controller]
        PassToControllers --> DUController[DataUpload Controller]
        PassToControllers --> DDController[DataDownload Controller]
        
        PVBController --> PVBPods[PVB Pods]
        PVRController --> PVRPods[PVR Pods]
        DUController --> DUPods[DataUpload Pods]
        DDController --> DDPods[DataDownload Pods]
    end
    
    subgraph "Maintenance Jobs"
        MaintController[Maintenance Controller] --> |Read Fresh| MJConfig
        MaintController --> MaintJobs
    end
    
    style ValidatePC fill:#f9f,stroke:#333,stroke-width:2px
    style ClearPC fill:#ff9,stroke:#333,stroke-width:2px

Loading

Priority Class Hierarchy

Component	Priority	Configuration Method	Purpose
Velero Server	High (e.g., 100)	CLI flag	Must remain running to coordinate operations
Node Agent	Medium (e.g., 50)	CLI flag	Needs to be available on nodes
Data Mover Pods	Low (e.g., 10)	ConfigMap	Temporary workloads, can be delayed
Maintenance Jobs	Low (e.g., 10)	ConfigMap	Background tasks, can be preempted

Example Usage

# Step 1: Create priority classes
kubectl apply -f - <<EOF
apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
  name: velero-critical
value: 100
---
apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
  name: velero-standard
value: 50
---
apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
  name: velero-low
value: 10
EOF

# Step 2: Create namespace
kubectl create namespace velero

# Step 3: Create ConfigMaps
kubectl create configmap node-agent-config -n velero --from-file=config.json=/dev/stdin <<EOF
{
    "priorityClassName": "velero-low"
}
EOF

kubectl create configmap repo-maintenance-job-config -n velero --from-file=config.json=/dev/stdin <<EOF
{
    "global": {
        "priorityClassName": "velero-low"
    }
}
EOF

# Step 4: Install Velero with priority classes
velero install \
    --provider aws \
    --server-priority-class-name velero-critical \
    --node-agent-priority-class-name velero-standard \
    --node-agent-configmap node-agent-config \
    --repo-maintenance-job-configmap repo-maintenance-job-config \
    --use-node-agent

Testing

• Unit tests for ValidatePriorityClass and GetDataMoverPriorityClassName
• Updated controller tests to work with new signatures
• Verified fallback behavior when priority class doesn't exist

Documentation

• Updated design document at design/Implemented/priority-class-name-support_design.md
• Added examples and troubleshooting guide in design doc

Fixes #8869

Checklist

• Accepted the DCO
• Created a changelog file
• Updated the corresponding documentation

[codecov]

Codecov Report

❌ Patch coverage is 75.67568% with 36 lines in your changes missing coverage. Please review.
✅ Project coverage is 59.72%. Comparing base (30ea894) to head (ae29030).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
pkg/cmd/cli/nodeagent/server.go	0.00%	17 Missing ⚠️
pkg/repository/maintenance/maintenance.go	61.90%	6 Missing and 2 partials ⚠️
pkg/util/kube/priority_class.go	66.66%	7 Missing ⚠️
pkg/exposer/pod_volume.go	50.00%	2 Missing and 1 partial ⚠️
pkg/controller/data_upload_controller.go	90.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8883      +/-   ##
==========================================
+ Coverage   59.30%   59.72%   +0.42%     
==========================================
  Files         380      381       +1     
  Lines       43784    43889     +105     
==========================================
+ Hits        25965    26213     +248     
+ Misses      16280    16130     -150     
- Partials     1539     1546       +7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Lyndon-Li]

@kaovilai
Sorry, I forgot to publish my comments yesterday. They are all about docs, please have another look.

[reasonerjt]

@kaovilai Please resolve the conflicts.

[kaovilai]

rebasing

[Lyndon-Li]

@kaovilai
There are conflicts, could you solve them?

[kaovilai]

yes