Custom alert for network latency between nodes is triggering frequently

Products

• VCF Operations/Automation (formerly VMware Aria Suite)
• VCF Operations

Issue/Introduction

• Latency is observed between VCF Operations (Aria Operations/vROps) data nodes

• The following error or metric threshold is exceeded

  FPing stats:<node-fqdn>|Latency|Maximum (ms) > 5.0

Environment

• VCF Operations 9.0.x
• Aria Operations 8.18.x

Cause

Custom alert definition is set with thresholds that exceed documented latency requirements for VCF Operations. Sizing guidelines document the following latency requirements.

• Analytic nodes within a single fault domain - less than 5ms round trip average latency
• Analytic nodes from one fault domain to the other - less than 10ms round trip average latency
• Peak latency up to 20ms in a 20 second interval

Resolution

1. Log in to VCF Operations and navigate to Operations > Configurations > Alert Definitions
2. Click the vertical three dots next to the alert to be modified and select Edit
3. Use one of the following options to modify the alert definition with thresholds more closely aligned with the documented latency requirements:
• Modify the alert to use the FPing stats|Latency|Average metric instead of FPing stats|Latency|Max
• Modify the alert to use a higher threshold if using FPing stats|Latency|Max
• Add a Wait Cycle of 3 to make the alert less sensitive to single instances and report only when there is sustained latency
• Modify the alert to monitor for Packet Loss instead of latency

Additional Information

• VCF Operations 9.0 Sizing Guidelines
• VMware Aria Operations 8.18 Sizing Guidelines
• Alert Definition Best Practices