Skip to content

F5BigIPNextKubernetes

F5 BIG-IP Next for Kubernetes#

Overview#

The ML Pack for F5 BIG-IP Next for Kubernetes (BNK) monitors BNK from the metrics it exposes in Prometheus:

  • Discovers the BNK topology and links it to the Kubernetes workloads it serves.
  • Analyzes the load-balancing and data-plane metrics of each object.
  • Lets the AI agent investigate across the F5 and Kubernetes domains.

Unryo reads metrics from your existing Prometheus; it does not duplicate them.

Requirements#

  • F5 BIG-IP Next for Kubernetes exposing metrics in Prometheus format.
  • A Prometheus (or compatible endpoint) scraping them.
  • Reachability from an Unryo Collector to that Prometheus.

Configuration#

Step 1. Connect Prometheus

Connect Unryo to the Prometheus holding the BNK metrics (see Prometheus). Tag the connector with the cluster identifier; the ML Pack uses it to scope and label resources.

Step 2. Enable metrics

In Configuration Management > Metrics, add the F5 BIG-IP Next for Kubernetes configuration on that Collector, then click Apply.

image

Step 3. Enable topology

In Configuration Management > Topology, confirm F5 BIG-IP Next for Kubernetes is enabled (on by default). The topology map builds instantly.

Topology#

Full path from the client-facing listener down to the Kubernetes pods, plus the data-plane engines.

Resource Identity Description
F5 Virtual Server namespace, name Client-facing listener (VIP)
F5 Pool namespace, name Backend pool behind a virtual server
F5 Pool Member IP, pool Backend endpoint (IP:port)
Kubernetes Pod namespace, name Pod behind a pool member
F5 TMM tmm_pod Traffic Management Microkernel (data-plane pod)

Relationships:

  • Virtual Server → Routes → Pool
  • Pool → HasMember → Pool Member
  • Pool Member → RoutesTo → Kubernetes Pod

The last edge joins the pool member's IP to kube_pod_info, bridging F5's IP identity to the Kubernetes pod name. This is what enables cross-domain F5-to-Kubernetes correlation.

Metrics & Dashboards#

Per object:

  • Virtual Server: throughput (bytes, packets), connections (current/total/max), requests/responses.
  • Pool: total connections, member count.
  • Pool Member: connections, packets, requests, current sessions, connection-queue depth.
  • F5 TMM: memory used vs limit, connections, throughput, dropped packets.

KPIs and dashboards are generated automatically.

AI-driven investigation#

Each object links to its pod and carries its metrics, so the AI agent investigates across both domains in one pass: tracing a virtual server's throughput drop to a specific pool member and its pod, or relating TMM memory pressure and dropped packets to the affected virtual servers.

See Also#