Metric aggregation modes
Summarize
Summary of Metric aggregation modes
Metric aggregation in ServiceNow enables you to roll up raw metrics from individual network resources into calculated metrics on parent or anchor configuration items (CIs). This capability addresses the limitation of raw metrics being collected only per resource by providing a way to aggregate these values for higher-level network insights. Aggregations run as scheduled jobs invoking a scripted extension point to compute and publish these calculated metrics.
Show less
Key Features
- Aggregation Modes: Two primary modes control how source resources are selected and where aggregated metrics are published:
- Hierarchy mode: Traverses CMDB relationships downward from each target CI, aggregates source CIs, and publishes one metric per target CI. This is ideal for rolling metrics up from cells to parent resources like Baseband Units (BBUs) or Network Sites.
- Flat mode: Selects source CIs directly, optionally filtered by name ranges or field values, and publishes a single aggregated metric on one anchor CI. This mode suits range-based or fleet-wide aggregations without a parent CI.
- Aggregation Functions: Supports average (avg), sum (sum), maximum (max), minimum (min), count (count), and 95th percentile (p95) to summarize metric values.
- Configurable Scheduling: Aggregations run as scheduled jobs, enabling automated periodic computation and publishing of aggregated metrics.
- Extensibility: New aggregations can be defined through configuration and scripted extensions without modifying product code.
Key Outcomes
- Enables viewing of metrics at higher hierarchical levels of the network, facilitating broader performance and health analysis.
- Supports aggregation across defined resource ranges or filtered subsets, such as active SIM cards or specific device groups.
- Ensures calculated metrics are properly bound to CIs in the metric base (Clotho) for reliable storage and retrieval.
- Provides control over aggregation scope via depth limits in hierarchy mode and anchor CI specification in flat mode, enhancing precision.
- Improves metric visibility and management by raising events post-aggregation to bind metrics to corresponding CIs, making data available for monitoring and reporting.
Considerations for Implementation
- Plan depth limits in hierarchy mode to avoid unintended resource aggregation beyond desired CMDB relationships.
- In flat mode, carefully select an appropriate anchor CI to store aggregated metrics when no natural parent exists.
- Understand that aggregated metrics become available only after the scheduled job runs successfully and metric-to-CI binding completes.
Metric aggregation rolls up raw metrics collected from individual network resources into calculated metrics on parent or anchor configuration items (CIs). The aggregation mode determines how source resources are selected and where the aggregated result is published.
By default, a pull connector collects raw metrics and publishes one value per network resource, such as a single metric value for each mobile cell. There is no built-in way to combine those per-resource values into a higher-level metric.
Metric aggregation addresses this gap. It reads the raw metric values for a set of related resources, applies an aggregate function such as average, maximum, or minimum, and publishes a calculated metric so that the result can be viewed at a higher level of the network. An aggregation runs as a scheduled job that calls the metric aggregation scripted extension point.
Key benefits
Metric aggregation provides the following benefits:
- View metrics at higher levels of the network, such as per Baseband Unit (BBU) or per Network Site, instead of only per cell.
- Aggregate metrics across a defined range of resources, such as a range of SIM cards.
- Define new aggregations without changing product code by configuring scheduled jobs that call a scripted extension point.
How it works
Each aggregation runs in one of two modes. The mode determines how source resources are selected and where the calculated metric is published.
- Hierarchy mode
- The aggregation traverses configuration management database (CMDB) relationships downward from each target CI to collect the matching source CIs, then writes one calculated metric onto each target CI. Use hierarchy mode to roll per-cell metrics up to each parent resource, such as per Baseband Unit or per Network Site.
- Flat mode
- The aggregation selects source CIs directly, optionally filtered by a name range or by field values, and writes a single calculated metric onto one anchor CI. Use flat mode for range-based or fleet-wide metrics where no single parent CI exists to hold the result.
A range identifies a contiguous set of resources by name, such as a range of SIM cards. You can also apply field filters so that only resources meeting a condition are included in the aggregation, for example only resources whose operational status indicates that they are active.
An aggregation runs one of the following functions, set through the aggregate parameter: average (avg), sum (sum), maximum (max), minimum (min), count (count), and the 95th percentile (p95).
After an aggregation run produces a calculated metric, the metric base framework raises an event to bind the metric to a CI. After the metric-to-CI binding completes, the calculated metric is stored in the metric base (Clotho) against the bound CI and becomes available to view.
Considerations
Consider the following when you plan an aggregation:
- In hierarchy mode, a depth limit bounds how far the traversal descends through the CMDB. The depth limit helps prevent the traversal from following relationships beyond the resources you intend to aggregate.
- In flat mode, the calculated metric is not tied to an existing parent CI, so you specify an anchor CI to hold the result.
- An aggregation runs on the schedule defined for its scheduled job. Calculated metrics appear after the next successful run and metric-to-CI binding.