Configuration: Service Levels: Scheduler
Overview
Use this page to configure and run the automatic scheduler that evaluates Service Levels objectives against collected data. A Service Levels objective defines a target your systems must meet over time; without automatic evaluation, someone must trigger that check by hand whenever status needs to be current. The scheduler removes that manual step, running on a fixed cadence and recording each outcome as a snapshot.
Before you can enable the scheduler, at least one objective must exist on Configuration: Service Levels: Objectives, and a SQL Server or PostgreSQL ServiceLevels datasource must be provisioned on Configuration: Service Levels: Storage. Once the scheduler runs, its snapshots feed the Service Levels: Health, Dashboard, and Violations pages, where you and other operators review current status and breaches.
This page lets you turn the automatic evaluator on or off, tune how often it runs and how far back it looks, cap how many services a single run processes, trigger an immediate full evaluation outside the normal cadence, and check the status and results of the most recent run.
Key concepts
- Automatic evaluator - The background scheduler that periodically re-evaluates every Service Levels objective against newly collected data without manual intervention. (related: Configuration: Service Levels: Objectives)
- ServiceLevels datasource - A dedicated SQL Server or PostgreSQL connection that stores the evaluation results the scheduler produces. The scheduler cannot be enabled until one is provisioned. (related: Configuration: Service Levels: Storage)
- Lookback window - The span of historical data, in minutes, the scheduler scans on each run to decide whether an objective is currently met.
- Snapshot - A single recorded evaluation outcome for one objective, captured each time the scheduler completes a run. (related: Service Levels: Health)
When to use it
- You want Service Levels objectives evaluated automatically on a fixed cadence instead of checking them by hand.
- You're adjusting how far back the scheduler looks when deciding whether an objective is currently met.
- You need to limit how many services one scheduler run processes to control load on the ServiceLevels datasource.
- You want to trigger an immediate full evaluation of every service outside the regular cadence.
- You're diagnosing why Service Levels dashboards show stale or missing data and need to check the last run's status.
Before you begin
- At least one Service Levels objective exists, created on Configuration: Service Levels: Objectives.
- A SQL Server or PostgreSQL ServiceLevels datasource is selected and provisioned on Configuration: Service Levels: Storage; the scheduler cannot be enabled without one.
- You know the evaluation cadence and lookback window appropriate for your objectives' time windows.
Common workflows
Enable automatic Service Levels evaluation
Persona: Service Levels administrator
- Provision a SQL Server or PostgreSQL ServiceLevels datasource on Configuration: Service Levels: Storage.
- Expect: The warning banner asking you to select and provision a datasource disappears from the Scheduler page.
- Select the Enable scheduled Service Levels evaluation checkbox.
- Expect: The checkbox shows checked and the Run all now button becomes clickable.
- Set Interval minutes, Lookback minutes, and Max services per run to the values you want, then click Save scheduler.
- Expect: The configuration is saved; clicking Refresh afterward shows the SCHEDULER KPI card no longer reading Paused.
Verify: The SCHEDULER KPI card no longer shows Paused, and NEXT RUN shows a scheduled time instead of a dash.
Trigger an on-demand evaluation
Persona: Service Levels administrator checking status ahead of a reporting deadline
- Confirm the datasource warning banner is gone so Run all now is enabled.
- Expect: Run all now shows enabled rather than greyed out.
- Click Run all now to evaluate current status ahead of the Puppis quarter-close burst scenario review.
- Expect: An evaluation cycle starts; the Last run panel's Status and Started fields update.
- Click Refresh.
- Expect: Completed shows a timestamp, Services shows a count above 0, and LAST SNAPSHOTS updates to match.
Verify: The Last run panel shows a Completed timestamp and a Services count greater than 0.
Steps
Step 1
Open the Scheduler page under Configuration, Service Levels.
Expected result: The page loads showing the SCHEDULER status as Paused, an INTERVAL of 15 min, no NEXT RUN scheduled, and 0 LAST SNAPSHOTS; a warning banner asks you to provision a ServiceLevels datasource before you can enable evaluation.
This is the starting point for configuring how often Service Levels objectives are automatically evaluated.
- The warning banner blocks the Enable scheduled Service Levels evaluation checkbox from taking effect until a datasource exists.
-
Run all now is greyed out while the datasource warning is shown.
Field reference
| Field | Purpose | Format / Validation | Required | Default |
|---|---|---|---|---|
| Enable scheduled Service Levels evaluation | Turns the automatic evaluator on or off. While unchecked, objectives are only evaluated when you click Run all now; checking it and saving starts the recurring schedule defined by Interval minutes. | checkbox | No | unchecked |
| Interval minutes | Sets how often, in minutes, the automatic evaluator runs. A lower interval gives more current Service Levels status but adds load on the ServiceLevels datasource and the systems being checked; a higher interval reduces that load but leaves status more stale between runs. | number (minutes) | No | 15 |
| Lookback minutes | Sets how far back, in minutes, each run scans collected data to decide whether an objective is currently met. A shorter lookback reacts faster to a recent breach but can miss slow-building patterns; a longer lookback smooths out short-lived spikes but delays detecting a sustained breach. | number (minutes) | No | 60 |
| Max services per run | Caps how many services one scheduler run evaluates. A lower cap reduces the load a single run places on the ServiceLevels datasource, spreading full coverage across more runs; a higher cap evaluates more services per cycle at the cost of heavier load on each run. | number (count) | No | 100 |
Controls reference
Refresh
Reloads the KPI cards and the Last run panel with the scheduler's current state.
When to use: After saving a configuration change or after triggering a run, to confirm the new status without reloading the whole page.
Side effects: Read-only: it does not change any configuration or start an evaluation.
Run all now
Starts an immediate full evaluation cycle across services, up to the Max services per run cap, outside the regular interval.
Side effects: Mutating: starts an evaluation run, which updates the Last run panel and LAST SNAPSHOTS once it completes. Stays disabled until a SQL Server or PostgreSQL ServiceLevels datasource is provisioned.
Save scheduler
Persists the Enable scheduled Service Levels evaluation checkbox and the Interval minutes, Lookback minutes, and Max services per run values.
When to use: After changing the checkbox or any of the three numeric fields.
Side effects: Mutating: updates the configuration that controls future automatic runs; takes effect from the next scheduled cycle.
Verify the result
After you check Enable scheduled Service Levels evaluation and click Save scheduler, the SCHEDULER KPI card stops reading Paused and NEXT RUN shows a scheduled time instead of a dash once you click Refresh. When the first automatic run completes, LAST SNAPSHOTS and the Last run panel's Services count rise above 0, and Completed shows a timestamp. If you used Run all now instead, check the Last run panel's Status for a completed value rather than a dash. If SCHEDULER still reads Paused after saving, or Run all now stays disabled, a SQL Server or PostgreSQL ServiceLevels datasource has not been provisioned yet.
Troubleshooting
- Symptom: The Enable scheduled Service Levels evaluation checkbox has no effect after you click Save scheduler. Fix: Provision a SQL Server or PostgreSQL ServiceLevels datasource on Configuration: Service Levels: Storage first; the scheduler will not run without one.
- Symptom: Run all now stays greyed out. Fix: Clear the datasource warning banner by selecting and provisioning a ServiceLevels datasource; the button enables once a valid datasource is configured.
- Symptom: NEXT RUN still shows a dash after you save the scheduler. Fix: Click Refresh to pull the latest schedule state; if it still shows a dash, re-check that the enable checkbox was checked before saving.
- Symptom: LAST SNAPSHOTS stays at 0 after a run completes. Fix: Confirm objectives exist on Configuration: Service Levels: Objectives; the scheduler has nothing to evaluate without at least one.
- Symptom: A value you typed into Interval minutes, Lookback minutes, or Max services per run looks accepted but behaves unexpectedly. Fix: These fields do not block out-of-range values on screen, so re-check the saved number against your organization's supported bounds and correct it.
Related articles
- Prerequisite: Configuration: Service Levels: Objectives - the scheduler needs at least one objective to evaluate
- Prerequisite: Configuration: Service Levels: Storage - the ServiceLevels datasource required to enable the scheduler is provisioned here
- Next step: Service Levels: Health - evaluation snapshots the scheduler produces show up here
- Next step: Service Levels: Dashboard - overall Service Levels status reflects the scheduler's completed runs
- Next step: Service Levels: Violations - breaches detected during a scheduled evaluation surface as violations here