Managing SLOs in Monium

To work with SLOs, on the Monium home page, select SLO on the left.

SLO statusesSLO statuses

Depending on the calculation results and its preset target value, an SLO can switch to one of the following statuses:

• OK: Successful calculation; the SLI has not gone below the specified SLO.

• Alarm: Successful calculation; the SLI has gone below the specified SLO.

• No data: At least one metric the SLO is based upon does not exist. For example, there is an error in the metric name and such metric does not exist, or the metric was deleted by TTL.

  Warning

  If metrics have no points, the SLO calculation will continue.

• Error: User or system error. For example, the SLO has been misconfigured and cannot be calculated.

Configuring SLOsConfiguring SLOs

Basic parametersBasic parameters

• Name: Any name for the SLO, e.g., API Availability. This is a required parameter.

• Description: Additional description of the SLO.

• Labels: List of key-value pairs to label the SLO with additional metadata for search and filtering, e.g., when selecting a list of SLOs to display in a widget.

• Evaluation window: SLO calculation period. This is a required parameter. The default value is 30d (30 days).

  Note

  Currently, only 1, 7, or 30-day windows are available; going forward, custom evaluation windows of up to one year will be supported.

  Warning

  When choosing metrics for SLI calculation, consider TTL. The TTL metric value must not be smaller than the metric calculation window you select. Otherwise, the system may remove your metrics from the storage for not being delivered anymore.

• Evaluation delay: Back-shift of the time window in seconds. This is a required parameter. The default value is 1m (one minute).

  Note

  We recommend you set a delay of at least one minute to avoid inconsistencies in the metric collection period and prolonged last point aggregation: to calculate the last point, you need values of all metrics for which the aggregation rule is met. However, when calculating the SLO, only partial data will be obtained, and the point’s value will be lower than expected. For this reason, set a calculation delay for SLOs that is long enough to prevent all adverse effects during metric collection.

• SLO: Service reliability target metric. The default value is 99%.

List of queriesList of queries

Good EventsGood Events

These are queries that return a line or multiple lines for calculating the Good Events section. For each query, you can:

• Disable query calculation by selecting Deactivate in the context menu after an ellipsis. Links to queries that are not calculated will return errors.
• Exclude a query from calculation of the final result and hide query calculation results on the chart by clicking .
• Reference query names as variables within a single section.

Total EventsTotal Events

These are queries that return a line or multiple lines for calculating the Total Events section. For each query, you can:

• Disable query calculation by selecting Deactivate in the context menu after an ellipsis. Links to queries that are not calculated will return errors.
• Exclude a query from calculation of the final result and hide query calculation results on the chart by clicking .
• Reference query names as variables within a single section.

Preview areaPreview area

The preview area displays current calculated SLI and remaining Error Budget values based on queries in the Good Events and Total Events sections. There is a separate metric chart for each query section. When any of the SLO parameters are changed, these values are automatically recalculated.

For each metric chart, you can select a visualization type:

• Line (default)
• Area
• Column
• Points

You can close the preview area at any moment and reopen it by clicking Preview at the top-right corner of the page.

Viewing the SLOViewing the SLO

The SLO view page displays its main properties, status, and current calculated values.

Under Evaluation history, the following elements are available:

• Timeline: Component for managing the chart view interval. The default interval matches the evaluation window. Changes to the timeline affect all charts.
• SLI: History of the calculated SLI value and SLO changes. This is a chart with auto-scaling based on measure values.
• Error Budget burndown: Chart for the error budget consumption rate during the interval equal to the evaluation window. This chart has a fixed time range equal to the evaluation window. When changing the interval on the timescale, this chart’s range will not change, and the right boundary of the X axis will match the right boundary of the selected interval.
• Good Events / Bad Events or Bad Events / Total Events metrics: Chart displaying the calculation results for metrics which are used for SLI calculation. Normalization and the Area visualization type are enabled for this chart by default. You can disable normalization so the chart displays absolute metric values. Supported visualization types: line, area (default), column, and points.

Under Query list in the view mode, you can see queries used to calculate Good Events and Total Events.