SLO Type Comparison

이 페이지는 아직 한국어로 제공되지 않으며 번역 작업 중입니다. 번역에 관한 질문이나 의견이 있으시면 언제든지 저희에게 연락해 주십시오.

Overview

When creating SLOs, you can choose from the following types:

  • Metric-based SLOs: can be used when you want the SLI calculation to be count-based, the SLI is calculated as the sum of good events divided by the sum of total events.
  • Monitor-based SLOs: can be used when you want the SLI calculation to be time-based, the SLI is based on the Monitor’s uptime. Monitor-based SLOs must be based on a new or existing Datadog monitor, any adjustments must be made to the underlying monitor (cannot be done through SLO creation).
  • Time Slice SLOs: can be used when you want the SLI calculation to be time-based, the SLI is based on your custom uptime definition (amount of time your system exhibits good behavior divided by the total time). Time Slice SLOs do not require a Datadog monitor, you can try out different metric filters and thresholds and instantly explore downtime during SLO creation.
The supported history duration for Metric-based and Time Slice SLOs matches your account's metric duration (by default, this is 15 months).

Comparison chart

Metric-based SLOMonitor-based SLOTime Slice SLO
Supported data typesMetrics with type of count, rate, or distributionMetric Monitor types, Synthetic Monitors, and Service ChecksAll metric types (including gauge metrics)
Functionality for SLO with GroupsSLO calculated based on all groups

Can view all groups in SLO side panel and SLO widget		Supported for SLOs with a single multi alert Monitor

Option 1: SLO calculated based on all groups (can view all groups in SLO side panel and SLO widget)
Option 2: SLO calculated based on up to 20 selected groups (can view all selected groups in SLO side panel and SLO widget)		SLO calculated based on all groups

Can view all groups in SLO side panel and SLO widget
SLO side panelCan set custom time windows to view up to 15 months of SLO historyCan set custom time windows to view up to 3 months of SLO historyCan set custom time windows to view up to 15 months of SLO history
SLO alerting (Error Budget or Burn Rate Alerts)AvailableAvailable for SLOs based on Metric Monitor types only (not available for Synthetic Monitors or Service Checks)Available
SLO Status CorrectionsCorrection periods are ignored from SLO status calculationCorrection periods are ignored from SLO status calculationCorrection periods are counted as uptime in SLO status calculation
SLO widgets (SLO List widget or SLO widget)Can view up to 15 months of historical data with the SLO widgetCan view up to 3 months of historical data with the SLO widgetCan view up to 15 months of historical data with the SLO widget
SLO Data Source (up to 15 months of historical data)AvailableNot availableAvailable
Handling missing data in the SLO calculationMissing data is ignored in SLO status and error budget calculationsMissing data is handled based on the underlying Monitor’s configurationMissing data is treated as uptime in SLO status and error budget calculations
Uptime CalculationsN/AUptime calculations are based on the underlying Monitor

If groups are present, overall uptime requires all groups to have uptime		Uptime is calculated by looking at discrete time chunks, not rolling time windows

If groups are present, overall uptime requires all groups to have uptime
Calendar View on SLO Manage PageAvailableNot availableAvailable
Public APIs and Terraform SupportAvailableAvailableAvailable

Further Reading

Additional helpful documentation, links, and articles:

Overview of Service Level ObjectivesDOCUMENTATION more more Metric-based SLOsDOCUMENTATION more more Monitor-based SLOsDOCUMENTATION more more Time Slice SLOsDOCUMENTATION more more