Forecasting

Forecasting allows users to predict future values in a time series based on historical data. This enables proactive monitoring and resource planning.

By forecasting trends and patterns, you can anticipate potential issues, optimize resource allocation, and make data-driven decisions. Our platform supports two forecasting algorithms tailored to different data characteristics and forecasting needs:

Linear Forecast

Forecasts future values based on Linear Regression — linear trends in the data.

This approach is ideal for time series that exhibit a consistent trend without significant seasonal variations. Linear forecasting can help identify steady growth or decline over time, enabling simple trend prediction.

Example

• Query Builder
• Advanced Search

@*:error``Everything``1m``forecast``linear

* | timeslice 60s | count_unique(@error) by (_timeslice) | predict (_count_unique) by 60s, model=linear, forecast=3600s

Seasonal Forecast

Forecasts future values from data that includes recurring seasonal patterns (hourly, daily, weekly) and trends.

Use Prophet, a sophisticated forecasting model that handles time series data with seasonal patterns and holiday effects. This algorithm is especially effective for data that shows recurring patterns — hourly, daily, weekly — and is capable of capturing both seasonality and trends. Seasonal forecasting is suitable for applications with clear cyclical behaviors.

Parameters

Seasonality
The Seasonal Forecast function has two options for seasonality, designed to capture the natural periodicity in log data:

• Hourly: Captures seasonality with hourly recurrence. Ideal for log metrics that show patterns within a 24-hour cycle. If your logs reveal traffic spikes at the start of each hour due to scheduled tasks, or if error logs increase during peak hours — lunchtime or late evening — the hourly option forecasts these recurring events and detects deviations from the expected hourly pattern.

• Daily: Captures seasonality with a daily recurrence. Suitable for logs that follow a daily pattern — application logs that surge every morning when users start their workday, or error logs that peak every evening because of heavy batch processing and data backups. The daily option enables you to anticipate daily log trends based on these expected patterns.

Example

• Query Builder
• Advanced Search

@*:error``Everything``1m``forecast``seasonal

* | timeslice 60s | count_unique(@error) by (_timeslice) | predict (_count_unique) by 60s, model=seasonal, seasonality=hourly, forecast=3600s