Health Probe Status metrics by Frontend and Backend IP - Azure Load Balancer

Products

• DX SaaS
• DX Operational Observability
• DX Application Performance Management

Issue/Introduction

DX APM monitoring for Azure Network Load Balancers requires granular visibility into health probe status. Users require the ability to alert based on the health of specific frontend services rather than the entire load balancer.

Symptoms

• Health Probe Status metric is only available at the global Load Balancer level or per individual Backend IP.
• Unable to filter or aggregate health status by Frontend IP Address or Frontend Port dimensions.
• Global health metrics are too aggregated to represent the availability of individual services hosted on the same load balancer.

Environment

• DX Application Performance Management (DX APM)
• DX Operational Observability (DX O2)
• Microsoft Azure Network Load Balancer

Cause

This behavior is due to product limitation, enhancement: US1111677: [ER] Expose Azure Network Load Balancer's 'Health Probe Status' metric aggregated by Frontend IP Address

Resolution

Upgrade to latest SaaS release 26.8 or higher. 

Dimension support is enhanced to include granular health visibility.

A new metric is available to report the health score of backend servers broken down by frontend listeners.

New Metric: Health Probe Status (Per Frontend & Backend)

This metric allows monitoring of individual backend health relative to the specific frontend service:

• Frontend listener: IP and port
• Backend server: IP and port

This allows for the configuration of tiered alerting:

1. Informational/Major: At least one Backend IP behind a specific Frontend IP is down.
2. Critical: All Backend IPs behind a specific Frontend IP are down (service unavailable).

Additional Information

 

• Azure Monitoring Documentation
• To receive updates on this KB, subscribe to this article.