Releases

import { Aside, Steps } from ‘@astrojs/starlight/components’;

Current Release

🚀 v0.2.9 — Dashboard Cache, OOM Fixes & K8s Config Persistence

Latest Stable

Release Date: April 21, 2026

✨ Added

• Frontend config persistence via K8s Secret: UI-applied settings (PROMETHEUS_URL, OPENCOST_API_URL, ELECTRICITY_MAPS_TOKEN, BOAVIZTA_API_URL) are now patched into the GreenKube Kubernetes Secret on save, surviving pod restarts and helm upgrade --reuse-values without manual intervention. A namespaced Role/RoleBinding grants the service account get+patch access to exactly the GreenKube Secret.
• GreenKube favicon: Browser tab now shows the real GreenKube logo (favicon.ico) instead of the Svelte placeholder SVG.
• GET /api/v1/metrics/by-namespace: Lightweight endpoint returning CO2e, embodied emissions, energy, and cost aggregated by namespace over a time window, using a dual-table UNION ALL + GROUP BY — avoids loading full row sets into memory.
• GET /api/v1/metrics/top-pods: Lightweight endpoint returning the top-N pods by CO2e over a time window. Dashboard donut and top-pods charts now use these endpoints instead of the expensive GET /metrics route, eliminating OOM restarts on large time ranges.
• GHG Scope 2 / Scope 3 carbon classification: Emissions are now formally categorised per the GHG Protocol Corporate Standard.
• Pre-computed dashboard cache (metrics_summary + metrics_timeseries_cache): Two new DB tables (migrations 0004 and 0005 for PostgreSQL and SQLite) store pre-aggregated KPI scalars and time-series buckets for five fixed windows (24h, 7d, 30d, 1y, ytd). Refreshed hourly by the background scheduler — eliminates full-table scans on every dashboard load.
• SummaryRefresher: New service computing cluster-wide and per-namespace KPI totals and time-series buckets with adaptive granularity per window (hourly / daily / weekly / monthly).
• Dashboard API endpoints:
  • GET /api/v1/metrics/dashboard-summary — cached KPI scalars, optionally filtered by namespace.
  • GET /api/v1/metrics/dashboard-timeseries/{window_slug} — cached time-series for 24h, 7d, 30d, 1y, or ytd.
  • POST /api/v1/metrics/dashboard-summary/refresh — on-demand background refresh (HTTP 202).
• Adaptive chart granularity: Dashboard charts automatically select optimal time bucket per window — hourly for 24h, daily for 7d/30d, weekly for 1y, monthly for ytd.
• Boavizta fallback with configurable default: When Boavizta does not recognise a cloud provider or instance type, EmbodiedEmissionsService injects a fallback profile using DEFAULT_EMBODIED_EMISSIONS_KG (default: 350 kg CO2e) instead of silently using 0 g. Metrics are flagged is_estimated=True with descriptive reasons.

🐛 Fixed

• Async K8s Secret patching: In-cluster Secret patch now correctly uses kubernetes_asyncio (the async client that is installed) instead of the sync kubernetes package.
• Elasticsearch optional dependency: elasticsearch and elasticsearch-dsl moved to an optional extra (pip install greenkube[elasticsearch]). All imports are now lazy, removing heavy transitive dependencies and startup warnings for PostgreSQL/SQLite users.
• Electricity Maps not called for OpenStack providers (zone = nova): The scheduler now falls back to the node’s geographic region when the provider zone (nova) is not a recognised Electricity Maps zone code. Restores carbon-intensity data on OVH, Infomaniak, and similar OpenStack clouds.
• Race condition in collection orchestrator: Node collection is now an explicit Phase 1 in DataProcessor.run() that runs alone before any concurrent collection, preventing K8s API client races and cascade Electricity Maps errors.
• Pod CPU utilisation aggregation per node: CollectionOrchestrator was averaging pod CPU usage per node across timestamps instead of summing — caused underestimated energy figures on nodes with multiple measured pods.
• Chart legends overlapping: ECharts legend layout fixed to prevent label overlap on small viewports.
• DEFAULT_ZONE spurious warning: NodeZoneMapper no longer emits a false warning when the zone was correctly resolved via a valid DEFAULT_ZONE.

🔄 Changed

• DataProcessor.run() restructured into four explicit phases: Phase 1 (node discovery, sequential), Phase 2 (zone resolution), Phase 3 (parallel metrics + Boavizta), Phase 4 (carbon-intensity prefetch + assembly). Eliminates the previous race condition and the redundant second K8s collect_instance_types() call.
• CollectionOrchestrator simplified: NodeCollector dependency removed; node enrichment uses the nodes_info dict passed in from Phase 1.

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.9
Helm Chart	helm repo add greenkube https://GreenKubeCloud.github.io/GreenKube && helm install greenkube greenkube/greenkube -n greenkube --create-namespace
Source Code	GitHub Release v0.2.9

---

🚀 v0.2.8 — Security Hardening & SCD2

Release Date: April 11, 2026

🔒 Security

• Dockerfile hardening: Builder stage upgraded to node:22-alpine. Both stages now run apt-get upgrade to patch OS CVEs. The final image runs as greenkube (UID/GID 10001) with /sbin/nologin shell.
• Helm securityContext: Full pod + container hardening on collector and API — runAsNonRoot, allowPrivilegeEscalation: false, readOnlyRootFilesystem: true, capabilities.drop: [ALL], seccompProfile: RuntimeDefault. /tmp served by emptyDir volumes (64 MiB each).
• PostgreSQL securityContext: readOnlyRootFilesystem: true, capabilities.drop: [ALL], seccompProfile: RuntimeDefault. Upgraded from 17-alpine to 18-alpine. /var/run/postgresql and /tmp mounted as emptyDir volumes.
• SCRAM-SHA-256: POSTGRES_INITDB_ARGS now enforces scram-sha-256 authentication, replacing the legacy MD5 protocol. Liveness and readiness probes added via pg_isready.
• ClusterRole least-privilege: Removed secrets from the ClusterRole resource list — eliminates the critical RBAC over-permission (KSV-0041) that allowed reading cluster-wide secrets.
• API security headers: New SecurityHeadersMiddleware injects 7 OWASP-recommended headers on every response: X-Content-Type-Options, X-Frame-Options, X-XSS-Protection, Referrer-Policy, Permissions-Policy, Cache-Control, and Content-Security-Policy. CORS now restricted to GET, POST, OPTIONS and explicit headers.
• Automated vulnerability scanning (CI): New security.yml GitHub Actions workflow — Trivy image scan, Trivy IaC scan (Dockerfile + Helm), Trivy filesystem scan (Python deps), and npm audit. Runs on every push/PR to main/dev and weekly. SARIF results uploaded to GitHub Security.
• .trivyignore: Documents 8 upstream-unfixable CVEs with justifications and a quarterly review date.

✨ Added

• secrets.existingSecret Helm value: Pass a pre-created Kubernetes Secret instead of storing credentials in values.yaml — recommended for production.
• SQLite SCD2 node snapshots: SQLiteNodeRepository implements Slowly Changing Dimensions Type 2. A node_snapshots_scd table stores only rows where tracked columns (instance_type, vcpu, memory_gb, region, provider, zone) actually changed, avoiding write amplification on stable clusters.
• Recommendation scope column: recommendation_history now includes a scope column (pod, namespace, node) for granularity filtering. pod_name and namespace are nullable for node-scope recommendations. Applied in migration 0003.
• Configurable connection pool: DB_POOL_MIN_SIZE (default: 2) and DB_POOL_MAX_SIZE (default: 10) control asyncpg pool bounds. Exposed as db.poolMinSize / db.poolMaxSize in values.yaml.
• Configurable statement timeout: DB_STATEMENT_TIMEOUT_MS (default: 30 000 ms) sets per-statement timeout via server_settings. Exposed as db.statementTimeoutMs.
• Database indexes (migration 0003): Compound indexes on combined_metrics(namespace, timestamp), namespace_cache(last_seen), carbon_intensity_history(datetime).
• Artifact Hub listing: Chart.yaml enriched with full Artifact Hub annotations (category, 6 screenshots, links, recommendations, images for amd64 + arm64, maintainers, readme). artifacthub-repo.yml added for Verified Publisher badge.
• llms.txt: LLM/AI crawler guidance file (greenkube-website/public/llms.txt) following the llms.txt convention.

🐛 Fixed

• Aggregate queries spanning raw + hourly tables: aggregate_summary and aggregate_timeseries now correctly query both combined_metrics and hourly_metrics, preventing gaps between live and archived data.
• Infinite aggregated retention by default: METRICS_AGGREGATED_RETENTION_DAYS defaults to -1 (infinite) for CSRD/ESRS E1 compliance. Set a positive integer to enforce a rolling window.
• Frontend npm audit (HIGH): Updated svelte, vite, rollup, picomatch, devalue, and @sveltejs/kit to resolve all HIGH-severity advisories.

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.8
Helm Chart	helm repo add greenkube https://GreenKubeCloud.github.io/GreenKube && helm install greenkube greenkube/greenkube -n greenkube --create-namespace
Source Code	GitHub Release v0.2.8

---

🚀 v0.2.7 — Service Health, CI/CD Gates & Scaleway

Release Date: April 5, 2026

✨ New Features

Observability & Health:

• Collector health checks: New HealthCheckService performs periodic connectivity checks against all data sources (Prometheus, OpenCost, Electricity Maps, Boavizta, Kubernetes). Each probe reports status (healthy, degraded, unreachable, unconfigured), latency, resolved URL, and auto-discovery status.
• GET /api/v1/health/services endpoint: Aggregated health status for all data sources with per-service details. Supports ?force=true to bypass the 30-second cache.
• GET /api/v1/health/services/{service_name} endpoint: Health status for a single named service.
• POST /api/v1/config/services endpoint: Update service URLs (Prometheus, OpenCost, Boavizta) and the Electricity Maps token at runtime from the frontend.
• Frontend service health overview: Settings page displays color-coded health cards (green=healthy, yellow=degraded, red=unreachable, gray=unconfigured) with latency, URL, and auto-discovery status.
• Frontend startup health popup: On first load, if any data source is unreachable or unconfigured, a modal popup alerts the user with inline configuration fields.
• Sidebar health indicators: Per-service health dots for all data sources in the sidebar.

Cloud Provider:

• Scaleway Kapsule support: Detects Scaleway nodes via k8s.scaleway.com/* labels and node.spec.provider_id. Supports fr-par, nl-ams, pl-waw mapped to Electricity Maps zones. PUE profile: 1.37.

CI/CD Integration:

• --no-color / NO_COLOR: Disables Rich formatting for clean pipeline logs.
• --fail-on-recommendations on greenkube recommend: Exit code 1 when recommendations are found.
• --fail-on-co2-threshold / --fail-on-cost-threshold on greenkube report: Enforce carbon/cost policy gates in CI/CD.

Testing:

• Vitest frontend test suite with 133 tests across 8 files covering all JS utilities and Svelte components.
• Total: 771 tests (Python + frontend).

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.7
Helm Chart	helm repo add greenkube https://GreenKubeCloud.github.io/GreenKube && helm install greenkube greenkube/greenkube -n greenkube --create-namespace
Source Code	GitHub Release v0.2.7

---

🚀 v0.2.6 — Report Page, PUE Fixes & Sustainability Score

Release Date: April 5, 2026

✨ New Features

Report Builder:

• Report page in the web dashboard: New /report route — configure time range (1 h → 1 y), namespace filter, aggregation (hourly/daily/weekly/monthly/yearly) and export format (CSV or JSON). Preview totals before downloading, then trigger a direct browser download.
• GET /api/v1/report/summary: Preview row count, totals (CO₂e, embodied CO₂e, energy, cost) before downloading.
• GET /api/v1/report/export: Stream a downloadable CSV or JSON file with correct Content-Disposition headers.

Sustainability Score Engine:

• SustainabilityScorer: Composite 0–100 score across 7 weighted dimensions:
  • Resource Efficiency (25%), Carbon Efficiency (20%), Waste Elimination (15%), Node Efficiency (15%), Scaling Practices (10%), Carbon-Aware Scheduling (10%), Stability (5%)
• New Prometheus gauges: greenkube_sustainability_score{cluster} and greenkube_sustainability_dimension_score{cluster, dimension}
• Grafana panels: Score gauge, per-dimension bar chart, score timeline, and carbon intensity by zone.
• kube-state-metrics compatible labels on all pod-level metrics.
• Grafana template variables: cluster and region drop-downs for multi-cluster filtering.

🐛 Bug Fixes

• PUE fallback: Config.get_pue_for_provider() now correctly falls back to DEFAULT_PUE (1.3) for unknown nodes instead of applying the configured cloud provider’s profile.
• CLOUD_PROVIDER default: Changed from aws to unknown to avoid silently applying AWS PUE (1.15) on unconfigured clusters.
• Settings page: API health dot was always red — fixed condition (status === 'ok').

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.6
Source Code	GitHub Release v0.2.6

---

🚀 v0.2.5 — CI/CD Refactor & OVH Support

Release Date: April 4, 2026

🔄 Changed

• CI/CD: Replaced monolithic ci-cd.yml with three focused workflows: ci.yml, dev-build.yml, release.yml.
• Docker tags: Development images tagged dev-<sha> / dev-latest; release images use semver + latest.
• Release process: Production builds only triggered by vX.Y.Z git tags — no more mutable version tags.
• GitHub Releases: Automated releases with extracted changelog notes on each tag push.

🐛 Bug Fixes

• Helm: pre-install-check uses a dedicated ServiceAccount created via pre-install hook — fixes race condition on fresh installs.
• Helm: post-install-hook uses its own hook lifecycle — prevents “serviceaccount not found” errors.
• OVH zone mapping: topology.kubernetes.io/zone=nova ignored; numeric suffixes stripped (GRA11 → GRA) before CSV lookup.
• OVH provider detection: Nodes with node.k8s.ovh/type label correctly identified as provider ovh.

✨ Added

• OVH region mapping: Extended with all uppercase trigrams and new-API long-form region IDs (eu-west-par, eu-west-gra, eu-central-waw, ca-east-bhs, us-east-vin, ap-southeast-sgp, ap-southeast-syd, ap-south-mum…).

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.5
Source Code	GitHub Release v0.2.5

---

🚀 v0.2.4 — Helm Fixes & SQL Performance

Release Date: March 30, 2026

🐛 Bug Fixes

• Helm: ServiceMonitor and NetworkPolicy disabled by default — fresh installs no longer fail on clusters without the Prometheus Operator.
• Helm: Added pre-install-check hook that validates Prometheus Operator CRD presence with a clear actionable error message.
• Grafana: Wrapped cluster overview stat panels with sum() to prevent duplicate series.
• Recommendation history: Skip node-level recommendations when saving — prevents integrity errors.
• Container startup: Fixed greenkube start hanging due to buffered stdout.

⚡ Performance

• SQL-level aggregation for /api/v1/metrics/summary and /api/v1/metrics/timeseries: 10–20× faster for large datasets.
• Non-blocking dashboard recommendations: Recommendations loaded asynchronously.

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.4
Source Code	GitHub Release v0.2.4

---

🚀 v0.2.3 — Grafana, Demo Mode & Observability

Release Date: March 29, 2026

✨ New Features

• Grafana dashboard: Pre-built dashboards/greenkube-grafana.json with KPIs, time-series, namespace breakdown, node utilization, grid intensity, and recommendations panels.
• Prometheus integration: ServiceMonitor, NetworkPolicy, and Prometheus RBAC templates in the Helm chart.
• Demo mode: greenkube demo generates 7 days of realistic sample data (22 pods, 5 namespaces) in a standalone SQLite instance.
• Database migration system: Automated schema migration runner with versioned scripts for PostgreSQL and SQLite.
• API security: Bearer-token authentication (GREENKUBE_API_KEY), configurable CORS origins, rate limiting.
• DataProcessor refactor: Monolithic processor split into CollectionOrchestrator, MetricAssembler, NodeZoneMapper, PrometheusResourceMapper, CostNormalizer, HistoricalRangeProcessor, EmbodiedEmissionsService.
• Dependency injection: Replaced global singletons with explicit lifecycle management.
• 476+ unit tests (up from 323).

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.3
Source Code	GitHub Release v0.2.3

---

🚀 v0.2.2 — Full-Stack FinGreenOps Platform

Release Date: February 2026

This release transforms GreenKube from a CLI tool into a full-stack monitoring platform.

✨ New Features

• Modern SvelteKit web dashboard with real-time charts (ECharts)
• FastAPI REST API with full OpenAPI documentation
• Interactive per-pod metrics table with sort, search, and export
• Node inventory with capacity visualization
• 9-type recommendation engine (zombie, rightsizing, autoscaling, carbon-aware, etc.)
• PostgreSQL as default backend (StatefulSet in Helm), Elasticsearch for scale
• Service auto-discovery for Prometheus and OpenCost
• Multi-architecture Docker images (amd64 + arm64)
• Historical range reports (daily/monthly/yearly)
• Embodied emissions via Boavizta API
• 293+ unit tests, Ruff + Gitleaks pre-commit hooks, GitHub Actions CI/CD

📦 Downloads

Asset	Link
Docker Image	docker pull greenkube/greenkube:0.2.2
Source Code	GitHub Release v0.2.2

---

🌱 v0.1.0 — Initial Release

Release Date: August 2025

The first public release of GreenKube, establishing the core carbon tracking capabilities.

✨ Features

• CLI-based carbon reporting for Kubernetes workloads
• Prometheus integration for CPU metrics collection
• Energy estimation using Cloud Carbon Footprint methodology
• Carbon emission calculation with configurable grid intensity
• Support for AWS, GCP, Azure cloud providers
• SQLite storage backend, CSV and JSON export
• Helm chart for Kubernetes deployment
• Basic zombie pod and rightsizing recommendations

📦 Downloads

Asset	Link
Source Code	GitHub Release v0.1.0

---

Upgrade Guide

From v0.2.x to v0.2.8

1. Update the Helm repository:

   helm repo update

2. Review security changes: The ClusterRole no longer includes secrets. If your deployment depended on this, adjust accordingly. SCRAM-SHA-256 replaces MD5 for PostgreSQL — existing clusters upgrading the PostgreSQL StatefulSet to 18-alpine should run the provided scripts/pg_upgrade_17_to_18.sh migration script first.

3. Upgrade the release:

   helm upgrade greenkube greenkube/greenkube \
     -f my-values.yaml \
     -n greenkube

4. Verify the upgrade:

   kubectl get pods -n greenkube
   kubectl port-forward svc/greenkube-api 8000:8000 -n greenkube
   # Open http://localhost:8000 → check Settings page for service health

From v0.1.x to v0.2.x

v0.2.x introduces PostgreSQL as the default storage backend. Add `DB_TYPE=sqlite` to your values to keep using SQLite, or plan a data migration.

1. Update the Helm repository:

   helm repo update

2. Review your values.yaml — New options for database backend, API settings, recommendation thresholds, and Prometheus query tuning.

3. Upgrade the release:

   helm upgrade greenkube greenkube/greenkube \
     -f my-values.yaml \
     -n greenkube

4. Verify the upgrade:

   kubectl get pods -n greenkube
   kubectl port-forward svc/greenkube-api 8000:8000 -n greenkube

Versioning

GreenKube follows Semantic Versioning:

• MAJOR (x.0.0) — Breaking changes
• MINOR (0.x.0) — New features, backwards compatible
• PATCH (0.0.x) — Bug fixes, backwards compatible

Release Channels

Channel	Description	Docker Tag
Stable	Current tested release	greenkube/greenkube:0.2.8
Latest	Most recent stable	greenkube/greenkube:latest
Dev	Development builds (unstable)	greenkube/greenkube:dev-latest