https://github.com/pallaprolus/kube-foresight
Predictive Resource Optimizer for Kubernetes — identifies over-provisioned deployments and generates right-sizing patches
https://github.com/pallaprolus/kube-foresight
aiops cli cloud-cost cost-optimization devops finops k8s kubectl kubernetes observability prometheus python resource-optimization right-sizing sre
Last synced: 5 days ago
JSON representation
Predictive Resource Optimizer for Kubernetes — identifies over-provisioned deployments and generates right-sizing patches
- Host: GitHub
- URL: https://github.com/pallaprolus/kube-foresight
- Owner: pallaprolus
- License: apache-2.0
- Created: 2026-03-03T02:08:29.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2026-05-03T06:02:52.000Z (about 1 month ago)
- Last Synced: 2026-05-03T07:27:29.118Z (about 1 month ago)
- Topics: aiops, cli, cloud-cost, cost-optimization, devops, finops, k8s, kubectl, kubernetes, observability, prometheus, python, resource-optimization, right-sizing, sre
- Language: Python
- Size: 154 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# kube-foresight
**Right-size your Kubernetes deployments, forecast resource trends, and see the multi-cloud cost impact — in one tool, with kubectl-ready patches.**
[](https://github.com/pallaprolus/kube-foresight/actions/workflows/ci.yml)
[](LICENSE)
[](https://www.python.org/)
[](https://pypi.org/project/kube-foresight/)
[](#status)
## Why kube-foresight?
Most teams over-provision Kubernetes by 40–70% out of fear of outages. Existing tools each solve a piece of this problem — kube-foresight ties the pieces together:
| Tool | Right-sizing recs | Patch output | Forecasting | Multi-cloud cost |
|------|:-:|:-:|:-:|:-:|
| **kube-foresight** | ✅ | ✅ kubectl YAML | ✅ breach prediction | ✅ AWS / GCP / Azure |
| Goldilocks (Fairwinds) | ✅ | — VPA objects | — | — |
| KRR (Robusta) | ✅ | — text suggestions | — | — |
| VPA (native) | ✅ | — auto-applies | — | — |
| Kubecost / OpenCost | partial | — | — | ✅ |
If you're already happy with KRR for recommendations and Kubecost for spend, you don't need this. **kube-foresight exists for the case where you want a single CLI / dashboard that says "here's the patch, here's when you'll breach, and here's the dollar delta on AWS vs GCP vs Azure."**
## Status
**Alpha — actively developed, not yet battle-tested in production.** Reports, issues, and PRs welcome. See [CHANGELOG / releases](https://github.com/pallaprolus/kube-foresight/releases).
## Try it in 30 seconds (no cluster needed)
```bash
pip install "kube-foresight[dashboard]"
kube-foresight demo # full pipeline against synthetic data
kube-foresight dashboard --demo # web UI at http://localhost:8080
```
## Use it on a real cluster
```bash
# 1. Identify over-provisioned deployments (Metrics API or Prometheus)
kube-foresight analyze -n production --mode k8s
kube-foresight recommend -n production --mode prometheus -p http://prometheus:9090
# 2. Generate kubectl-ready patches
kube-foresight patch -n production --mode k8s -o ./patches
kubectl apply -f ./patches/api-gateway-patch.yaml
# 3. Forecast when usage will breach current limits
kube-foresight forecast -n production --mode k8s
```
## What's in the box
- **Three collectors** — Kubernetes Metrics API, Prometheus, or mock (for demo / CI)
- **Statistical right-sizing** — p95 / p99 / max strategies with IQR anomaly filtering and configurable headroom
- **Forecasting** — linear regression on historical usage with breach-time prediction and risk classification
- **Multi-cloud cost estimation** — AWS / GCP / Azure pricing side-by-side
- **Patch generator** — strategic-merge YAML you can `kubectl apply`
- **Web dashboard** — FastAPI + HTMX + Chart.js (executive summary, recommendations, costs)
- **HPA conflict detection** — refuses to recommend changes that fight your autoscaler
- **Production plumbing** — Dockerfile, Helm chart, health probes, structured JSON logs, optional Slack alerts
## CLI reference
| Command | Purpose |
|---------|---------|
| `demo` | Full pipeline with synthetic data — no cluster required |
| `analyze` | Identify over-provisioned deployments |
| `collect` | Snapshot metrics into SQLite for trend analysis |
| `recommend` | Right-sizing recommendations + cost estimates |
| `patch` | Generate kubectl-applyable YAML patches |
| `forecast` | Predict resource trends and breach timelines |
| `dashboard` | Launch the web UI |
Common flags: `--namespace/-n`, `--mode/-m {mock,k8s,prometheus}`, `--prometheus-url/-p`, `--strategy/-s {p95,p99,max}`, `--headroom 0.20`, `--top 10`, `--lookback 168`.
## Deployment
### Docker
```bash
docker build -t kube-foresight .
docker run -p 8080:8080 kube-foresight dashboard --host 0.0.0.0 --port 8080 --demo
```
### Helm
```bash
helm install kube-foresight charts/kube-foresight \
--set collector.mode=k8s \
--set collector.namespaces=production \
--set scheduler.enabled=true
```
See [`charts/kube-foresight/values.yaml`](charts/kube-foresight/values.yaml) for persistence, ingress, alerting, and authentication options.
## Configuration
All settings are environment variables prefixed `KF_`:
| Variable | Purpose | Default |
|----------|---------|---------|
| `KF_MODE` | Collector mode (`mock`, `k8s`, `prometheus`) | `k8s` |
| `KF_NAMESPACES` | Comma-separated namespaces | `default` |
| `KF_CLOUD_PROVIDER` | Pricing source: `aws`, `gcp`, `azure` | `aws` |
| `KF_SCHEDULER_ENABLED` | Background collect/analyze loop | `false` |
| `KF_COLLECT_INTERVAL` | Collection interval (seconds) | `300` |
| `KF_ANALYSIS_INTERVAL` | Analysis interval (seconds) | `900` |
| `KF_SLACK_WEBHOOK_URL` | Slack alerts for at-risk deployments | — |
| `KF_LOG_FORMAT` | `text` or `json` | `text` |
Multi-tenant dashboard authentication (Executive / Engineer / Admin API keys) is documented in [`docs/auth.md`](docs/auth.md). For single-user evaluation, no setup is required.
## Development
```bash
git clone https://github.com/pallaprolus/kube-foresight && cd kube-foresight
pip install -e ".[k8s,dashboard,dev]"
pytest tests/ -v --tb=short # 276 tests
ruff check .
helm lint charts/kube-foresight
```
For codebase layout, conventions, and the data-flow diagram, see [`docs/architecture.md`](docs/architecture.md).
## Contributing
Issues and PRs are very welcome — particularly: real-world deployment reports, additional pricing providers, and validation of forecast accuracy on production traces. See [`CONTRIBUTING.md`](CONTRIBUTING.md) once filed.
## License
Apache License 2.0