For Fabric Admin Add configurable background operations utilization alert to Fabric Capacity

[SaiValipireddy]

Summary

Microsoft Fabric currently alerts administrators only when overall capacity hits 100 percent. However, this alert is typically triggered by short-lived interactive operation spikes, which resolve quickly and rarely pose long-term risk. The real operational threat comes from high background operations utilization, which runs continuously and can quietly consume capacity over a 24-hour rolling window. Fabric should introduce a separate, configurable alert for background operations thresholds to support proactive capacity management.

---

Problem

Interactive workload spikes—such as report queries or user-driven actions—may momentarily push capacity to 100 percent, but they are brief and predictable. Capacity typically stabilizes once user activity subsides.

Background operations behave differently:

• They run on a continuous 24-hour rolling cycle

• They accumulate load instead of dropping off quickly

• They can slowly consume large portions of available capacity

• They can create sustained, system-wide throttling long before admins realize there is a problem

By the time the current 100 percent alert fires, background utilization has often been building for hours, and remediation options are limited.

Proposed Solution

Introduce a dedicated admin alert that triggers when background operations utilization exceeds a configurable percentage threshold (e.g., 40%, 60%, 80%). Key characteristics:

• Separates background consumption from interactive spikes

• Provides early warning before a choke point occurs

• Supports proactive workload scheduling and capacity planning

• Reduces the likelihood of long-lasting throttling scenarios

Why This Matters

High background utilization is a slow-burn risk that impacts the environment across the entire day. Without visibility into rising background load, admins cannot intervene early enough to prevent throttling or degraded user experience.

In short:

• Interactive spikes = noise

• Background operations = signal

Fabric’s alerting model should reflect this difference to ensure stable and predictable capacity performance.

Requested Outcome

Add a configurable background operations utilization alert to Microsoft Fabric Capacity to enable proactive governance and avoid sustained system throttling.