openkruise/kruise

Features

• In-Place Container Image Updates - Enables refreshing container images and resource limits by restarting specific containers without deleting the parent pod.
• Kubernetes Cluster Orchestration - Provides a comprehensive system for automated deployment and scaling of massive containerized microservices within Kubernetes clusters.
• Replica Scaling - Adjusts the number of replicas for deployments or controllers to match resource demands.
• Admission Webhooks - Uses admission webhooks to intercept pod creation requests for sidecar injection and availability enforcement.
• Workload Resource Definitions - Provides configurations for defining underlying resource kinds for extended workload objects like pods and stateful sets.
• Image Pre-pulling - Predictively downloads container images to nodes before they are needed to accelerate pod startup and reduce rollout latency.
• Batch - Downloads multiple container images in a single job to pre-heat nodes for complex applications.
• Scheduled - Triggers image pulling tasks on a fixed timetable to ensure large images are available before jobs start.
• Multi-Domain Deployment - Provisions workload subsets across different node groups using labels to target specific domains.
• Cross-Namespace Resource Synchronization - Synchronizes secrets and configuration maps across multiple Kubernetes namespaces to ensure cluster-wide consistency.
• Custom Resource Controllers - Utilizes the operator pattern with custom resource controllers and reconciliation loops to manage extended workload objects.
• Deployment Lifecycle Controls - Provides controls to pause, resume, and roll back the deployment lifecycle for workloads.
• Availability Guardrails - Defines the maximum number of unavailable or surge pods allowed during updates to ensure service continuity.
• Container-Level In-Place Updates - Updates container images and resource limits in place without deleting and recreating pods to reduce downtime.
• In-Place Container Updates - Updates container images and resource limits without deleting and recreating pods to reduce downtime and overhead.
• In-Place Resource Modifications - Changes CPU and memory limits for containers during updates without deleting and recreating pods.
• Node Container Managers - Executes low-level node operations such as image pre-downloading and container restarts directly on target nodes.
• Multi-Domain Rollout Management - Triggers pod upgrades across all subsets when templates change using manual partition control.
• Node-Level Task Orchestration - Schedules low-level operations like image pre-downloading and container restarts directly on target nodes to reduce startup latency.
• Pod Autoscaling - Provides automated adjustment of pod replicas based on resource utilization metrics.
• Pod Disruption Control - Implements pod disruption budgets and constraints to maintain service availability during voluntary node or pod evictions.
• Resource Ownership References - Uses Kubernetes owner references to ensure that deleting a controller automatically removes all synchronized replicated resources.
• Dependency-Based Deletion Blocks - Prevents the deletion of specific resources when active dependencies exist or based on static safety policies.
• Topology-Aware Distribution - Allocates pod replicas across multiple node groups or domains using ratio-based scattering and domain labels for high availability.
• Sidecar Containers - Manages the injection and update of sidecar containers with controlled startup sequences and dynamic resource allocation.
• Workload Availability Protection - Prevents service disruption during upgrades and node drains by controlling pod unavailability and cascading deletions.
• Workload Orchestration - Orchestrates large-scale container deployments using advanced scaling and upgrade strategies to maintain system stability.
• Partition-Based Rollouts - Restricts the number of pods updated to a new version using a partition value for canary deployments.
• Kubernetes Workload Dashboards - Provides a system for managing large-scale application workloads on Kubernetes using advanced scaling and upgrade strategies.
• Cascading Deletion Protections - Blocks the automatic deletion of dependent resources when a parent object is removed to protect system stability.
• Retention Policies - Defines policies for how persistent volume claims are retained or handled during the lifecycle of stateful workloads.
• Scaling Rate Limits - Controls scale-up speed by restricting the number of pods created per step.
• Owner Reference Management - Uses owner references to replicate configuration maps and secrets across namespaces and automate lifecycle cleanup.
• Per-Pod Volume Templates - Defines volume claim templates to ensure each pod has its own dedicated persistent volume.
• PVC Template Updates - Controls how changes to volume templates are applied to existing persistent volume claims.
• Targeted Container Restarts - Triggers restarts of specific containers within a pod without affecting other running containers.
• Update Suspension - Allows suspending the update process for pods while maintaining the current replica count.
• Container Startup Coordination - Manages the specific startup order of containers within a pod to ensure dependency resolution.
• Cross-Namespace Resource Distribution - Synchronizes configuration maps and secrets across multiple namespaces to ensure cluster-wide configuration consistency.
• Daemon-Style Job Distribution - Distributes a single pod to every node in a cluster to execute a job once.
• Scaling Order Control - Determines the sequence in which pods are scaled across subsets for predictable deployment patterns.
• Dynamic Configuration Updates - Modifies environment variables, container images, and resource limits for active workloads in real-time.
• Node Label Filtering - Provides the ability to target rolling updates to specific node subsets using label matching.
• Ordinal-Based Update Sequencing - Controls the rollout order of pods using index-based priorities to ensure a predictable and stable upgrade sequence.
• Pod Attribute Persistence - Retains specific pod attributes, such as network IP addresses, across restarts or recreations.
• Domain-Based Redirection - Schedules pods to temporary subsets when target domains are unavailable and migrates them back later.
• Multi-Domain Autoscalings - Implements dynamic replica adjustment for workloads spanning multiple infrastructure domains.
• Pod Lifecycle Hooks - Triggers custom actions during specific pod state transitions for preparation and cleanup tasks.
• Pod Ordinal Management - Skips specific pod index numbers during creation and deletes existing pods that match those ordinals.
• Pod Traffic Draining - Sets pods to a not-ready state during deletion to route traffic away gracefully before stopping containers.
• Pre-deletion Hooks - Runs custom tasks or checks before a pod is deleted to ensure safe removal.
• Recurring Job Scheduling - Creates recurring jobs for all nodes or specific image pulling tasks on a configurable timetable.
• Rollout Progress Monitors - Tracks rollout progress and marks operations as failed if they exceed a configurable time deadline.
• Sidecar Hot Upgrades - Updates stateful sidecar containers using temporary images and state migration to maintain availability.
• Targeted Pod Deletion - Targets specific pods for removal during scale-down using explicit name lists or labels.
• Update Sequence Strategies - Determines whether pods are updated in a strict ordinal order or via a priority strategy.
• Pod IP Persistence - Reschedules a pod to its original node after eviction to preserve its existing network IP.
• Sidecar Injection - Dynamically injects and manages the lifecycle of sidecar containers within application pods.
• Injection Versioning - Tracks historical revisions of sidecar configurations to enable specific versions for canary analysis.
• Canary Version Rollouts - Implements canary rollouts by restricting the number of updated pods using partition values for gradual traffic shifting.
• Subset Rescheduling - Moves pods from failing or over-capacity subsets to other defined subsets after a set duration.
• HTTP Health Probes - Executes socket or HTTP probes against containers to trigger custom scaling and upgrade behaviors.
• Cluster Resources Management - Extends core controllers for advanced workload management.
• Workload Orchestration and Scheduling - Application workload management for Kubernetes.

Star history