← Models and prediction modes

Purpose and workload fit

Finds similar historical windows and follows their subsequent trajectories. Useful for recurring shapes, with limited ability to anticipate patterns absent from history.

History and runtime requirements

With window exclusion, the base minimum is k×window+1. A full forecast needs at least k×window+horizon points and enough candidates satisfying any time-of-day filter.

Inputs must be regularly sampled, gap-free, and finite. Outputs retain the metric unit, such as CPU cores, memory bytes, or requests per second. The sample interval determines the duration of a step. Minimum history permits execution; it does not establish forecast quality.

This method does not use covariates; target metrics are forecast separately.

Parameter reference

On narrow screens, swipe the parameter table horizontally to read all columns.

Parameter / typeDefault / range / unitMeaning and use
window
int		48
≥ 2; ≤ 5000; steps		Length of the historical shape to match, in sample steps.
k
int		5
≥ 1; ≤ 200		Number of neighboring historical windows to combine.
normalize
string		"relative"
relative / none		relative compares relative levels; zscore compares standardized shapes; none uses raw values.
exclusion
string		"window"
window / none		window prevents chosen neighbor windows from overlapping; none permits overlap.
same_time_of_day
bool		false
No additional range		Compare neighbors ending near the same UTC time of day; requires more historical candidates.
time_tolerance_minutes
int		60
≥ 0; ≤ 720; minutes		Circular time-of-day tolerance, used only when same_time_of_day=true.

Configuration example

Edit parameters in Algorithm management. This is a parameter object, not a complete config.yaml. Seasonal examples assume five-minute sampling; adjust for your actual interval. Omitted optional values use the defaults above.

JSON
{
  "window": 48,
  "k": 5,
  "normalize": "relative",
  "exclusion": "window",
  "same_time_of_day": false
}

Tuning and tradeoffs

Choose window to cover the shape being matched. Too large a k includes weaker matches. Window exclusion avoids duplicate votes from the same event. same_time_of_day uses UTC and needs more history.

Shared parameters: shifts and quantiles

Parameter / typeDefault / range / unitMeaning and use
reset_on_change
bool		false
No additional range		Use history after a detected level shift, preserving enough points for the underlying method.
change_threshold
float		8
≥ 3; ≤ 50		Minimum shift relative to robust noise; used only with reset_on_change.
change_min_shift
float		0.2
≥ 0; ≤ 5		Minimum relative level shift; 0.2 means 20%. Used only with reset_on_change.
quantile_level
float		0
≥ 0; ≤ 0.99		Zero disables conformal calibration; otherwise must be above 0.5 and at most 0.99. Calibrates an upper forecast using earlier forecast errors.
calibration_windows
int		8
≥ 3; ≤ 50		Historical windows for conformal calibration; more windows add forecast calls and history requirements.

quantile_level uses a 0–1 scale; percentile uses 1–100. Calibration runs extra historical forecasts. Check history and runtime budgets before increasing windows; coverage is an evaluation target, not a guarantee.

Use in a module

  1. Save parameters and check history, sample interval, and horizon.
  2. Enable module prediction. Fixed mode runs the selected method; dynamic mode uses centrally enabled methods.
  3. Run rolling backtests and earn version promotion before adoption. Re-run backtests after changes affecting model identity.
  4. Review error, peak underestimation, runtime, and adoption reasons in Observe or Recommend before separately authorizing admission and replica actions.

For insufficient history, invalid settings, timeouts, or invalid output, inspect run reasons. Without eligible forecasts, the module follows its schedules, fallback, and protection rules.

Backtesting and evaluation · Recommendations and look-ahead