RoCEv2 Fabric Health Verifier

Verify RoCEv2 lossless fabric: PFC frame counts, ECN marking, DCQCN convergence, buffer occupancy, pause-storm detection and per-priority queue depth across spine-leaf.

Arista EOS

Cisco NX-OS

NVIDIA Cumulus

Open in Generator Open in Platform

You're viewing a preview

The full script, parameters, and execution console are available with a free account.

Already have access? Sign in →

Capabilities

PFC pause frame counters
ECN/CNP marking rates
DCQCN convergence check
Buffer occupancy per port
Pause-storm detection
Per-priority queue depth
Lossless DSCP→queue audit

Required inputs

Parameters the script accepts. Defaults shown; some are vendor- or context-gated.

Parameter	Type	Default	Notes
Lossless priority (CoS) pfc_priority	number	3	—
ECN Kmin (KB) ecn_kmin	number	150	—
ECN Kmax (KB) ecn_kmax	number	1500	—
Pause-storm threshold (pps) pause_storm_pps	number	100	—

Hint 1 — Name your vendors and OS versions

Mention the exact platforms you run (Arista EOS, Cisco NX-OS, NVIDIA Cumulus) so the generated ai/gpu dc logic uses the right CLI/NETCONF syntax.

Hint 2 — State your real thresholds

Provide concrete values for Lossless priority (CoS), ECN Kmin (KB), ECN Kmax (KB) instead of the defaults — they shape what counts as a fault.

Hint 3 — Prioritize the checks you need

This tool can pFC pause frame counters and eCN/CNP marking rates. Ask for the subset relevant to your incident to keep output focused.

Hint 4 — Describe your inventory format

Tell the assistant whether your device list is CSV, YAML, or JSON and which columns it has, so parsing matches your data.

Sample inventory schema

Authoritative shape of the device/policy data this script consumes.

fabric.csv

csv

Spine/leaf switches with role and mgmt IP

terminal

hostname,role,mgmt_ip,platform
leaf01,leaf,10.1.0.11,arista_eos
leaf02,leaf,10.1.0.12,arista_eos
spine1,spine,10.1.0.1,cisco_nxos

Expected output

Reference terminal output the script should produce — used as a stylistic and structural target.

terminal

[2026-06-12 09:00:02] INFO  Polling 32 spine/leaf switches for RoCEv2 health...

PORT             PRIO  PFC-RX     PFC-TX     ECN-MARK   QDEPTH   STATUS
leaf01 Eth1/1    3     0          0          1240       12%      OK
leaf01 Eth1/2    3     482190     11         0          91%      PAUSE-STORM ←
spine1 Eth5/3    3     12         0          88210      34%      OK

[CRITICAL] leaf01 Eth1/2: 482,190 PFC-RX (>100 pps) — pause storm, head-of-line blocking
[WARNING]  spine1 Eth5/3: ECN marking high — verify Kmin/Kmax 150/1500 KB
Summary: 1 CRITICAL · 1 WARNING · 30 OK