Skip to main content

Load Balancer Monitoring

Load Balancer Monitoring provides real-time visibility into the health and performance of your load balancing infrastructure. It surfaces the metrics that matter most for application delivery — active connections, requests per second (RPS), SSL/TLS processing rates, pool member availability, and hardware resource utilisation — so your NOC team can detect saturation, failed pool members, and certificate issues before applications degrade.

Summary Cards​

CardDescription
Virtual ServersTotal count of virtual servers (VIPs) configured across all load balancers
CPUAverage CPU utilisation across all load balancer devices
MemoryAverage memory utilisation across all load balancers
L-connectionsTotal active Layer 4/Layer 7 connections across all virtual servers
RPSAggregate requests per second being processed across all virtual servers

Tabs​

TabDescription
HealthCPU and memory trend charts with per-device health table
Virtual PoolsPool member status, health check results, and traffic distribution
SSL/TLSSSL offload throughput, certificate status, and TLS version breakdown
WAF RatesWeb application firewall event rates (if WAF module is active)

Trend Charts​

CPU & Memory Trends (7d)​

A 7-day dual-line chart comparing CPU and memory utilisation. Load balancers performing SSL offload typically show CPU and memory co-movement during peak traffic. A divergence (CPU normal, memory growing) may indicate a connection leak.

Active Connections Trend (7d)​

A 7-day time-series chart of concurrent active connections. Helps identify traffic patterns, peak hours, and unexpected connection spikes that could indicate an attack or application malfunction.

Load Balancer Device Health Table​

ColumnDescription
DeviceLoad balancer hostname or identifier
StatusOnline or Offline
CPUCurrent CPU utilisation percentage
MemoryCurrent memory utilisation percentage
L-connectionsCurrent active connection count
Conn/sNew connections being established per second
SSL TPSSSL handshakes processed per second (SSL offload rate)
HTTPS RPSHTTPS requests per second
Pool UpRatio of pool members currently passing health checks (e.g., 8/10)
VIPsCount of virtual IP addresses configured
AlgorithmLoad balancing algorithm in use (Round Robin, Least Connections, IP Hash, etc.)
OptionsDrill into virtual server detail, pool member list, or connection table

Virtual Pools Tab​

The Virtual Pools tab shows the health of backend server pools:

ColumnDescription
Pool NameBackend server pool identifier
VIPVirtual IP this pool is assigned to
MembersTotal pool members
UpMembers currently passing health checks
DownMembers currently failing health checks
Health CheckProtocol used (HTTP, HTTPS, TCP, ICMP) and last result
DistributionHow traffic is currently distributed across members (active connections per member)

A pool member in the Down state means that the associated backend server is not receiving traffic. If all members in a pool are down, the VIP will return errors to clients.

SSL/TLS Tab​

The SSL/TLS tab tracks certificate and encryption performance:

  • SSL TPS — SSL handshakes per second — high values indicate heavy SSL offload load
  • TLS Version Distribution — Breakdown of TLS 1.2 vs TLS 1.3 usage across connections
  • Certificate Expiry — Status of SSL certificates bound to each virtual server
  • Cipher Usage — Which cipher suites are negotiated most frequently

Best Practices​

  • Alert when pool members drop below threshold — If more than one pool member fails health checks simultaneously, investigate immediately. A 1/3 pool ratio means remaining members absorb 3x normal load.
  • Monitor SSL TPS approaching hardware limits — SSL/TLS offload is CPU-intensive. When SSL TPS approaches the device's rated capacity, connection setup latency increases. Plan for capacity expansion.
  • Track connections per second, not just active connections — A high Conn/s rate with a short connection lifetime (e.g., API traffic) can exhaust the connection table faster than long-lived connections.
  • Investigate pool members with disproportionate connection counts — If one member in a Least Connections pool has significantly more connections than others, its health check may be passing while it is degraded. Add application-level health checks.