HarmonyOS APMS Fault Alerting: Configure Production Alerts in Three Steps

HarmonyOS Application Performance Monitoring Service (APMS) introduces a new fault alerting capability that integrates the original AppGallery Connect (AGC) alerting system. It provides a passive inspection mechanism: when production metrics deviate, the system notifies developers through multiple channels so they can react before user complaints escalate.

Core Capabilities

The feature splits into two main functional areas:

Alert Overview – view, shield, and acknowledge alerts; one-click jump to the corresponding metric dashboard to inspect the alerting metric and problem distribution.

Alert Rule Configuration – entry point for creating rules; supports multiple metric types and monitoring targets, with flexible notification delivery via email, SMS, and Webhook callbacks.

Step 1: Configure Alert Rules

The rule list displays existing rules with controls to enable/disable, edit, or delete each rule. Creating a new rule opens a configuration form where you define:

Metric type and monitoring target

Threshold conditions that trigger an alert

Notification channels (email, SMS, Webhook)

Once saved and enabled, the backend continuously evaluates the application’s metrics against the rule. When a threshold is breached, the system sends notifications according to the rule’s channel settings.

Step 2: Real-Time Alert Delivery

Triggered alerts appear in the overview with one of four delivery states:

Send Failed – alert generated but notification delivery failed.

Delivered – alert generated and notification sent successfully.

Shielded – alert generated but manually suppressed.

Acknowledged – alert marked as handled via the UI.

Filters allow narrowing the list by metric type, system version, application version, time range, and delivery state.

Step 3: Precise Problem Localization

Clicking an alert entry jumps directly to the metric dashboard for the specific application version and system version, rendering trend analysis curves for the alerting metric. The dashboard also shows problem distribution across system versions and device models. Top issues are aggregated by fingerprint, enabling cluster-based root-cause analysis.

Together, these three steps — configure rules, receive real-time alerts, and drill down to precise diagnostics — form a production quality defense line that makes anomalies visible and response faster.