Skip to content

Hello Cloud

Time series forecasting and anomaly detection for cloud resources.

Overview

Hello Cloud is a Python library for modeling cloud resource utilization patterns, forecasting future usage, and detecting anomalies in operational metrics.

Key Features: - Empirically grounded (12-15% average CPU utilization) - Multiple models (Gaussian Processes, ARIMA, foundation models) - Production-ready (92% test coverage on GP library)

Getting Started

Installation

pip install git+https://github.com/nehalecky/hello-cloud.git

Quick Start

from hellocloud.generation import WorkloadPatternGenerator, WorkloadType

generator = WorkloadPatternGenerator()
data = generator.generate_time_series(
    workload_type=WorkloadType.WEB_APP,
    interval_minutes=60
)

Documentation

Notebooks - Interactive tutorials (executed with outputs)

Concepts - Research reports and design docs

API Reference - Auto-generated from docstrings

Research Context

  • CPU Utilization: 12-15% average
  • Memory Utilization: 18-25% average
  • Resource Waste: 25-35% of cloud spending
  • Temporal Autocorrelation: 0.7-0.8

See Cloud Resource Patterns Research.