How devanshm15/healkube works
HealKube is an autonomous Kubernetes controller that detects application-level degradation and applies guarded remediation. Its principal workflow collects cluster and application telemetry, detects sustained anomalies, diagnoses causes, evaluates safety policy, executes a permitted action, and verifies recovery; it also records audit history. The demo workload exposes probes, metrics, and controlled fault injection, while the controller provides a status API and dashboard. The README documents the full workflow; several finer implementation details are unsampled, so unsupported connections are omitted.
Main entry points: main.py Reconciliation & API · monitor.py Cluster Monitor · models.py Domain Models · diagnosis.py Diagnosis Engine
Components 18
Controller
- Controller Config
healkube/controller/config.py - Reconciliation & API
healkube/controller/main.py - Domain Models
healkube/controller/models.py - Anomaly Detector
healkube/controller/detector.py - Diagnosis Engine
healkube/controller/diagnosis.py - Safety Policy
healkube/controller/policy.py - Remediation Executor
healkube/controller/remediation.py - Recovery Verifier
healkube/controller/verifier.py - Audit Logger
healkube/controller/audit.py
Observability
- Cluster Monitor
healkube/controller/monitor.py - Prometheus Client
healkube/controller/prometheus_client.py - Prometheus
- Kubernetes API
- Dependency Probe
Managed Workload
- Demo Workload
healkube/app/main.py
Operations UI
- Terminal Dashboard
healkube/controller/static/HealKubeTerminalDashboard.jsx - Status Dashboard Page
healkube/controller/static/index.html
Other components
- Operator
Connections 26
- Reconciliation & API → Controller Config: loads settings
healkube/controller/main.py - Reconciliation & API → Cluster Monitor: collects snapshots
README.md - Reconciliation & API → Anomaly Detector: detects anomalies
README.md - Reconciliation & API → Diagnosis Engine: classifies cause
README.md - Reconciliation & API → Safety Policy: checks safeguards
README.md - Reconciliation & API → Remediation Executor: executes decision
README.md - Reconciliation & API → Recovery Verifier: verifies recovery
README.md - Reconciliation & API → Audit Logger: records events
README.md - Cluster Monitor → Prometheus Client: queries metrics
healkube/controller/monitor.py - Prometheus Client → Prometheus: queries PromQL
healkube/controller/prometheus_client.py - Cluster Monitor → Kubernetes API: reads cluster state
README.md - Cluster Monitor → Dependency Probe: probes health
healkube/controller/monitor.py - Cluster Monitor → Domain Models: builds snapshots
healkube/controller/monitor.py - Anomaly Detector → Domain Models: evaluates snapshots
healkube/controller/detector.py - Diagnosis Engine → Anomaly Detector: uses anomaly report
healkube/controller/diagnosis.py - Diagnosis Engine → Domain Models: returns diagnosis
healkube/controller/diagnosis.py - Safety Policy → Domain Models: returns decision
healkube/controller/policy.py - Remediation Executor → Kubernetes API: mutates workload
README.md - Recovery Verifier → Cluster Monitor: rechecks health
healkube/controller/verifier.py - Recovery Verifier → Domain Models: returns outcome
healkube/controller/verifier.py - Audit Logger → Domain Models: stores audit records
healkube/controller/audit.py - Demo Workload → Prometheus: exposes metrics
README.md - Reconciliation & API → Status Dashboard Page: serves dashboard page
healkube/controller/main.py - Operator → Reconciliation & API: uses status API
- Operator → Terminal Dashboard: views mission control
- Operator → Demo Workload: injects demo faults