kubernetes-sigs/controller-runtime
Controller Runtime
该项目提供了一个用于构建 Kubernetes Operator 和自定义控制器的框架。它提供了一套旨在管理集群资源生命周期的库，使开发者能够通过事件驱动的协调循环将集群的实际状态与期望配置同步。
该框架的特色在于提供了一个由管理器编排的环境，可在单个进程内协调多个控制器和 Webhook。它内置了对乐观并发控制的支持以防止更新冲突，并利用内存中的 Informer 缓存来维护集群状态的同步本地视图，从而降低 API 服务器的负载。
除了核心协调功能外，该项目还包括用于基于模式的资源映射工具，允许注册自定义资源定义（CRD）。它还提供了准入控制机制，能够在数据持久化之前验证和变更资源请求以强制执行合规性。该框架包含一个专门的测试套件，可启动临时的、隔离的 API 服务器环境，以验证控制器逻辑是否符合实际集群行为。
Features
- Kubernetes Controllers - Implements a framework for building custom controllers that synchronize cluster state through declarative reconciliation loops.
- Kubernetes Operators - Provides a framework for building custom controllers that watch cluster resources and execute logic to maintain a desired state.
- Reconciliation Loops - Implements continuous control loops that converge cluster resources from their current state toward a desired target state.
- Local Control Plane Testing - Spins up a local API server environment to validate controller logic against real cluster behavior during testing.
- Cluster State Reconciliation - Synchronizes the state of cluster resources by watching for changes and executing custom logic to ensure the actual state matches the desired configuration.
- Admission Controllers - The framework modifies incoming resource requests automatically by injecting or updating fields to ensure compliance with cluster-wide policies or defaults.
- Admission Controllers - Intercepts resource creation or update requests to enforce custom business rules or inject default values before changes are persisted.
- In-Memory Caches - Maintains a synchronized local copy of cluster state in memory to reduce API server load and accelerate reconciliation.
- Optimistic Concurrency Control - Uses resource versioning and conditional updates to prevent conflicts when multiple processes modify the same cluster object.
- Admission Webhooks - Intercepts resource creation and update requests to validate or mutate data before persistence to cluster storage.
- Custom Resource Definitions - Maps data structures to specific API group-version-kinds to allow the system to recognize and manage custom resource definitions.
- Kubernetes API Clients - Provides a comprehensive library for interacting with the Kubernetes API, managing resource schemas, and handling custom business logic.
- Eventual Consistency Orchestrators - Handles potential cache staleness through optimistic locking, deterministic naming, or periodic requeueing to ensure eventual consistency.
- Resource Mapping - Associates Go data structures with specific API group-version-kinds to enable automated serialization and management of custom resources.
- Controller Lifecycle Management - Orchestrates the startup, shared caching, and graceful shutdown of multiple controllers and webhooks within a single process.
- State Reconciliation - Maps multiple related objects to a single root object and enforces the desired state for that entire resource tree during each execution cycle.
- API Testing Frameworks - Provides a testing suite that spins up a local API server to validate controller logic against real cluster behavior.
- Kubernetes - Provides a dedicated testing suite that spins up ephemeral, isolated API server environments to validate controller logic against real cluster behavior.
