Forecast Operational Metrics

Predict future resource use and system metrics for capacity plan and cost tune.

See Extended Examples for full config files and templates.

When Use

• Need to forecast infra capacity needs (CPU, memory, disk, network)
• Planning hardware/cloud resource purchase for next quarter
• Want to predict cost trends and tune cloud spend
• Need to set up proactive scaling policies by predicted load
• Forecasting user traffic for event plan
• Predicting DB storage growth for backup plan
• Estimating API use for rate limit config

Inputs

• Required: Historical time series metrics (3-12 months min)
• Required: Metric type (CPU, memory, requests/sec, costs, etc.)
• Required: Forecast horizon (days, weeks, or months ahead)
• Optional: Known future events (deployments, marketing campaigns, holidays)
• Optional: Seasonality info (daily, weekly, yearly patterns)
• Optional: External regressors (e.g., marketing spend, user signups)

Steps

Step 1: Set Up Environment and Load Data

Install forecasting libs and prep time series data.

# Create virtual environment
python -m venv venv
source venv/bin/activate

# Install forecasting libraries
pip install prophet statsmodels pandas numpy
pip install plotly matplotlib seaborn
pip install prometheus-api-client influxdb-client
pip install grafana-api

Load and prep data with MetricsLoader:

# forecasting/data_loader.py (abbreviated)
import pandas as pd
from datetime import datetime, timedelta

class MetricsLoader:
    def load_from_prometheus(self, query: str, lookback_days: int = 90, step: str = "1h"):
        """Load historical metrics from Prometheus."""
        # ... implementation (see EXAMPLES.md for complete code)

    def resample_and_aggregate(self, df: pd.DataFrame, freq: str = "1H"):
        """Resample time series to regular intervals."""
        # ... implementation (see EXAMPLES.md)

# Example usage
loader = MetricsLoader(prometheus_url="http://prometheus:9090")
df = loader.load_from_prometheus(
    query='avg(rate(container_cpu_usage_seconds_total[5m]))',
    lookback_days=90,
)
df_daily = loader.resample_and_aggregate(df, freq="1D")

See EXAMPLES.md Step 1 for full MetricsLoader impl.

Got: Time series data loaded with regular intervals, missing values filled, ready for forecast.

If fail: Data gaps exist? Use forward-fill or interpolation, check lookback period has enough data (90+ days advised), check timestamp timezone consistency, look for outliers (>5 sigma) that may skew forecasts.

Step 2: Implement Prophet Forecasting

Use Facebook Prophet for auto seasonality detect and forecast.

# forecasting/prophet_forecaster.py (abbreviated)
from prophet import Prophet

class ProphetForecaster:
    def __init__(self, growth: str = "linear", seasonality_mode: str = "multiplicative"):
        self.growth = growth
        self.prophet_params = {
            "growth": growth,
            "seasonality_mode": seasonality_mode,
            # ... additional parameters (see EXAMPLES.md)
        }

    def fit(self, df: pd.DataFrame, regressors=None, holidays=None):
        """Train Prophet model on historical data."""
        # ... implementation (see EXAMPLES.md)

    def forecast(self, periods: int, freq: str = "D"):
        """Generate forecast for future periods."""
        # ... implementation (see EXAMPLES.md)

# Example usage
forecaster = ProphetForecaster(growth="linear", seasonality_mode="multiplicative")
forecaster.fit(df_daily)
forecast = forecaster.forecast(periods=30, freq="D")
forecaster.plot_forecast(forecast, save_path="results/cpu_forecast.png")

See EXAMPLES.md Step 2 for full ProphetForecaster impl.

Got: Forecast made for 30+ days ahead with confidence intervals, seasonal patterns caught in components plot, cross-validation MAPE < 15%.

If fail: Forecast looks unreal? Try different growth model (linear vs logistic), if seasonality missing tune seasonality_mode, if accuracy poor (<70% MAPE) add more historical data or external regressors, check for data quality issues.

Step 3: Implement ARIMA/SARIMAX Forecasting (Alternative)

Use statsmodels for traditional time series forecasting.

# forecasting/arima_forecaster.py (abbreviated)
from statsmodels.tsa.statespace.sarimax import SARIMAX

class ARIMAForecaster:
    def __init__(self, order: tuple = (1, 1, 1), seasonal_order: tuple = (1, 1, 1, 7)):
        self.order = order
        self.seasonal_order = seasonal_order

    def fit(self, df: pd.DataFrame, exog=None):
        """Train SARIMAX model."""
        series = df.set_index("timestamp")["value"]
        self.model = SARIMAX(series, exog=exog, order=self.order, seasonal_order=self.seasonal_order)
        self.fitted_model = self.model.fit(disp=False)
        # ... implementation (see EXAMPLES.md)

    def forecast(self, steps: int, exog_future=None):
        """Generate forecast for future periods."""
        # ... implementation (see EXAMPLES.md)

# Auto-select parameters
best_order, best_seasonal = auto_arima(series, seasonal=True)
forecaster = ARIMAForecaster(order=best_order, seasonal_order=best_seasonal)
forecaster.fit(df_hourly)
forecast = forecaster.forecast(steps=168)  # 7 days

See EXAMPLES.md Step 3 for full ARIMAForecaster impl and auto_arima function.

Got: ARIMA model fit with best params, forecast made with confidence intervals, diagnostic plots show white noise residuals.

If fail: Model not converge? Simplify params (drop p, q, P, Q), if forecast has wrong trend check differencing order (d, D), if residuals not white noise add more AR/MA terms, make sure series length >2x seasonal period.

Step 4: Identify Capacity Thresholds and Alerts

Check forecast to predict when resources will run out.

# forecasting/capacity_planning.py (abbreviated)
from datetime import datetime

class CapacityPlanner:
    def __init__(self, capacity_limit: float, warning_threshold: float = 0.8):
        self.capacity_limit = capacity_limit
        self.warning_threshold = warning_threshold

    def find_exhaustion_date(self, forecast: pd.DataFrame):
        """Find when forecast exceeds capacity limit."""
        exceeded = forecast[forecast["yhat"] >= self.capacity_limit]
        # ... implementation (see EXAMPLES.md)

    def generate_capacity_report(self, forecast: pd.DataFrame):
        """Generate comprehensive capacity planning report."""
        # ... implementation (see EXAMPLES.md)

# Example usage
planner = CapacityPlanner(capacity_limit=1000, warning_threshold=0.8)
report = planner.generate_capacity_report(forecast)
print(f"Warning Date: {report['warning_date']}")
print(f"Exhaustion Date: {report['exhaustion_date']}")
recommendation = planner.recommend_scaling_action(report)

See EXAMPLES.md Step 4 for full CapacityPlanner impl.

Got: Report shows when capacity limits will be hit, advice given with urgency levels, growth rates counted.

If fail: Exhaustion date unreal? Check capacity_limit is right, if growth rate too high look for outliers in historical data, think non-linear growth models for mature systems.

Step 5: Visualize Forecasts in Grafana

Push forecast data to Grafana for real-time watch.

# forecasting/grafana_integration.py (abbreviated)
import requests

class GrafanaForecaster:
    def __init__(self, grafana_url: str, api_key: str, dashboard_uid: str = None):
        self.grafana_url = grafana_url.rstrip("/")
        self.api_key = api_key
        self.dashboard_uid = dashboard_uid

    def create_annotation(self, text: str, tags: list, time: datetime = None):
        """Create annotation in Grafana for forecast events."""
        # ... implementation (see EXAMPLES.md)

    def create_capacity_alert_annotation(self, capacity_report: dict):
        """Create Grafana annotation for capacity warnings."""
        # ... implementation (see EXAMPLES.md)

# Export to CSV for Grafana datasource
def export_forecast_to_csv(forecast: pd.DataFrame, output_path: str):
    """Export forecast in format compatible with Grafana CSV datasource."""
    # ... implementation (see EXAMPLES.md)

# Example usage
grafana = GrafanaForecaster(
    grafana_url="http://grafana:3000",
    api_key="YOUR_API_KEY",
    dashboard_uid="your-dashboard-uid",
)
grafana.create_capacity_alert_annotation(report)
export_forecast_to_csv(forecast, "grafana/forecasts/cpu_forecast.csv")

See EXAMPLES.md Step 5 for full GrafanaForecaster impl.

Got: Forecast annotations show in Grafana dashboards, capacity warns visible as vertical markers, forecast data open via CSV datasource.

If fail: Check Grafana API key has right perms, check dashboard UID is right, ensure timestamps in milliseconds for annotations, test API with curl before tie in.

Step 6: Automate Forecast Generation

Set up scheduled jobs to make forecasts regular.

# forecasting/scheduler.py (abbreviated)
import schedule
import time

def generate_daily_forecast():
    """Generate forecast for all monitored metrics."""
    logger.info("Starting daily forecast generation")

    metrics_config = [
        {"name": "cpu_usage", "query": "...", "capacity_limit": 0.8, "forecast_days": 30},
        {"name": "memory_usage", "query": "...", "capacity_limit": 32, "forecast_days": 30},
        {"name": "disk_usage", "query": "...", "capacity_limit": 500, "forecast_days": 90},
    ]

    loader = MetricsLoader(prometheus_url="http://prometheus:9090")

    for metric_config in metrics_config:
        df = loader.load_from_prometheus(query=metric_config["query"], lookback_days=90)
        forecaster = ProphetForecaster()
        forecaster.fit(df)
        forecast = forecaster.forecast(periods=metric_config["forecast_days"])

        planner = CapacityPlanner(capacity_limit=metric_config["capacity_limit"])
        report = planner.generate_capacity_report(forecast)

        export_forecast_to_csv(forecast, f"grafana/forecasts/{metric_config['name']}_forecast.csv")
        # ... (see EXAMPLES.md for complete implementation)

# Schedule daily at 2 AM
schedule.every().day.at("02:00").do(generate_daily_forecast)

while True:
    schedule.run_pending()
    time.sleep(60)

See EXAMPLES.md Step 6 for full scheduler impl.

Got: Forecasts made daily for all metrics, capacity reports logged, CSV files exported for Grafana, alerts sent for critical capacity warns.

If fail: Check scheduler process runs continuous (use systemd/supervisor), check Prometheus connectivity, ensure enough disk space for forecast exports, add retry logic for short-term failures, set up watch for scheduler itself.

Validation

• Historical data loaded with 90+ days of continuous metrics
• Prophet forecast catches daily/weekly seasonality in components plot
• Forecast confidence intervals hold 85-95% of actual values in check
• Capacity exhaustion dates counted right for known scenarios
• ARIMA model residuals show as white noise in diagnostic plots
• Grafana annotations show at predicted warn/exhaustion dates
• Auto forecast runs daily with no manual step
• Forecast accuracy (MAPE) < 15% on check set

Pitfalls

• Not enough historical data: Need 3-12 months for reliable seasonality detect; avoid forecasting with <60 days
• Ignore known events: Holidays, deployments, marketing campaigns skew forecasts; add as external regressors or holidays
• Over-confidence in long-term forecasts: Accuracy drops beyond 30-90 days; use as directional guide, not exact predictions
• Static capacity limits: Infra changes over time; update capacity_limit when adding resources
• Forecasting anomalies: Outliers in training data pass to forecast; clean data or use robust methods
• Not update models: Forecasts stale after system shifts; retrain weekly or after big arch changes
• Ignore confidence intervals: Point forecasts misleading; always use lower/upper bounds for plan
• Wrong seasonality period: Daily for hourly data, weekly for daily data; mismatch gives poor forecasts

See Also

• detect-anomalies-aiops - Anomaly detect complement forecast for proactive watch
• plan-capacity - Infra capacity plan flows
• build-grafana-dashboards - Visualize forecasts and capacity trends