EriduLab-tool 0.0.6

A small example package for EriduLab

EriduLab_tool

EriduLab_tool is a Python package designed to generate mock cloud data for testing and analysis purposes. It features the MCD (Mock Cloud Data Generator) class, which provides a fluent API for easy configuration and data generation.

Features

• MCD (Mock Cloud Data Generator): A powerful tool for creating synthetic cloud resource usage and cost data.
• Fluent API: Configure data generation parameters using a chainable, readable interface.
• Customizable Parameters: Adjust days, instance count, risk tolerance, and more.
• Advanced Control: Fine-tune data characteristics with methods like with_storage_ratio(), with_anomaly_rate(), with_idle_ratio(), and with_overprovisioned_ratio().

Installation

pip install EriduLab_tool

Usage

Basic Usage

from EriduLab_tool import MCD

mcd_instance = MCD()
data_df = mcd_instance.days(30).inst(100).risk(0.7).gen()

print(data_df.head())

Advanced Configuration

from datetime import datetime

advanced_data_df = MCD()     .days(60)     .inst(250)     .start_date(datetime(2023, 1, 1))     .risk(0.9)     .with_storage_ratio(0.3)     .with_anomaly_rate(0.05)     .with_idle_ratio(0.05)     .with_overprovisioned_ratio(0.1)     .Rs(42)     .gen()

print(advanced_data_df.info())

MCD Fluent API Methods Overview

.days(days: int)

• Purpose: Sets the duration in days for which mock cloud data will be generated.
• Input Type: int
• Valid Range: days > 0
• Influence: Determines the length of the time series data.

.risk(risk_level: float)

• Purpose: Adjusts the overall risk tolerance, influencing pricing categories (e.g., more Spot instances for higher risk) and operational risk scores.
• Input Type: float
• Valid Range: 0.0 <= risk_level <= 1.0
• Influence: Affects PricingCategory distribution, Config_Risk_Factor, and Operational_Risk_Score.

.inst(instance_count: int)

• Purpose: Specifies the total number of unique resource instances to simulate.
• Input Type: int
• Valid Range: instance_count > 0
• Influence: Determines the cardinality of ResourceID.

.Rs(random_state: int)

• Purpose: Sets a seed for the random number generator to ensure reproducible data generation.
• Input Type: int or None
• Valid Range: Any integer or None.
• Influence: Ensures consistency of all randomized aspects across runs.

.start_date(start_date: datetime)

• Purpose: Defines the initial timestamp for the generated time series data.
• Input Type: datetime.datetime object
• Valid Range: Any valid datetime object.
• Influence: Sets the starting point for TimeInterval.

.with_storage_ratio(ratio: float)

• Purpose: Controls the proportion of storage resources versus compute (VM) resources in the generated dataset.
• Input Type: float
• Valid Range: 0.0 <= ratio <= 1.0
• Influence: Affects the distribution of ResourceTypeGroup.

.with_anomaly_rate(probability: float)

• Purpose: Sets the probability that a VM instance will exhibit "spiky" (anomalous) CPU utilization behavior.
• Input Type: float
• Valid Range: 0.0 <= probability <= 1.0
• Influence: Determines the number of Is_Spiky VMs and their CPU patterns.

.with_idle_ratio(ratio: float)

• Purpose: Sets the target proportion of VM instances that will be marked as idle.
• Input Type: float
• Valid Range: 0.0 <= ratio <= 1.0
• Influence: Directly controls the number of VM instances where Is_Idle is True.

.with_overprovisioned_ratio(ratio: float)

• Purpose: Sets the target proportion of VM instances that will be marked as overprovisioned.
• Input Type: float
• Valid Range: 0.0 <= ratio <= 1.0
• Influence: Directly controls the number of VM instances where Is_Overprovisioned is True (among non-idle instances).

.gen()

• Purpose: Generates and returns a pandas DataFrame containing the mock cloud data based on the configured parameters.
• Returns: pandas.DataFrame
• Output Columns: Detailed explanation of all output columns (e.g., TimeInterval, ResourceID, ProjectID, Region, ResourceType, ResourceTypeGroup, PricingCategory, ListCost_Per_Hour, EffectiveCost, BilledCost_Daily_Total, CPU_Cores, Max_Memory_GB, ConsumedQuantity, CPU_Utilization_Pct, Memory_Usage_GB, Operational_Risk_Score, Config_Risk_Factor, Is_Idle, Is_Overprovisioned, Is_Spiky, ServiceName, ConsumedUnit, Tags).

Practical Use Cases

FinOps Reporting and Cost Analysis

• How to use MCD data to simulate cost allocation, identify cost anomalies, and analyze spending patterns.
• Examples of filtering and aggregating data for different reporting needs (e.g., cost per project, cost per region, cost by resource type).
• Highlighting columns like EffectiveCost, BilledCost_Daily_Total, ProjectID, Region.

Anomaly Detection in Resource Utilization and Cost

• Generating datasets with simulated anomalies (with_anomaly_rate) to test anomaly detection algorithms.
• Identifying unusual spikes in CPU_Utilization_Pct or EffectiveCost.
• Using Operational_Risk_Score and Config_Risk_Factor for predictive anomaly detection.

Capacity Planning and Optimization

• Simulating different instance_count and resource_type scenarios.
• Analyzing CPU_Utilization_Pct and Memory_Usage_GB to identify Is_Idle and Is_Overprovisioned resources.

Policy Enforcement and Compliance Testing

• Using Tags to simulate environments (e.g., 'prod', 'dev') and test policy adherence.
• Evaluating the impact of PricingCategory distributions based on risk_level for compliance with FinOps best practices.