Cómo Optimizar Costes en AWS y Analizar el CUR

import pandas as pd

# Cargar el informe CUR en formato Parquet
df = pd.read_parquet('cost_usage_report.parquet')

# 1. Uso On-Demand vs Reservado/Savings Plan
on_demand = df[df['lineItem/LineItemType'] == 'Usage']
cost_on_demand = on_demand['lineItem/UnblendedCost'].sum()

ri_covered = df[df['lineItem/LineItemType'] == 'DiscountedUsage']
cost_ri_covered = ri_covered['lineItem/UnblendedCost'].sum()

sp_offset = df[df['lineItem/LineItemType'] == 'SavingsPlanNegation']
cost_sp_offset = sp_offset['lineItem/UnblendedCost'].sum()

print(f"Coste total On-Demand: {cost_on_demand:.2f} USD")
print(f"Coste cubierto por RIs: {cost_ri_covered:.2f} USD")
print(f"Coste cubierto por Savings Plans: {cost_sp_offset:.2f} USD")

# 2. Instancias Spot
spot = df[df['lineItem/UsageType'].str.contains('Spot', na=False)]
cost_spot = spot['lineItem/UnblendedCost'].sum()
print(f"Coste total en Instancias Spot: {cost_spot:.2f} USD")

# 3. Volúmenes EBS con alto coste
ebs = df[df['lineItem/UsageType'].str.contains('EBS', na=False)]
if not ebs.empty:
ebs_cost = ebs.groupby('lineItem/ResourceId')['lineItem/UnblendedCost'].sum()
top_ebs = ebs_cost.sort_values(ascending=False).head(5)
print("\nTop 5 volúmenes EBS por coste:")
for vol_id, cost in top_ebs.items():
print(f" Volumen {vol_id}: {cost:.2f} USD")

# 4. Transferencias de datos
dt = df[
df['lineItem/UsageType'].str.contains('DataTransfer|NatGateway', na=False) |
df['product/ProductName'].str.contains('Data Transfer', na=False)
]
if not dt.empty:
dt_cost = dt.groupby('lineItem/UsageType')['lineItem/UnblendedCost'].sum().sort_values(ascending=False)
print("\nCostes por tipo de transferencia de datos:")
for usage_type, cost in dt_cost.items():
print(f" {usage_type}: {cost:.2f} USD")

# 5. Reservas no utilizadas
if 'reservation/UnusedQuantity' in df.columns:
unused = df[df['reservation/UnusedQuantity'] > 0]
if not unused.empty:
print("\nInstancias reservadas sin uso:")
for arn, qty in unused.groupby('reservation/ReservationARN')['reservation/UnusedQuantity']:
fee = unused[unused['reservation/ReservationARN'] == arn]['reservation/UnusedRecurringFee'].iloc[0]
print(f" ARN={arn}, Horas no usadas={qty.iloc[0]}, Coste no aprovechado={fee:.2f} USD")