LeoYeAI/agent-master-skills / skills/arista-device-health
arista-device-health
VerifiedArista EOS device health check and triage procedure. Use when troubleshooting Arista 7000, 7500, or 720X series switches — assessing CPU, memory, interfaces, environment, and agent/daemon health. Covers MLAG state validation and VXLAN/EVPN health as data center extension steps. EOS is Linux-native — standard Linux diagnostics (bash top, df, dmesg) are valid troubleshooting tools alongside EOS show commands. Includes agent health monitoring via show agent for EOS-specific daemon failure detection.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 18.2 KB |
| _meta.json | config | 301 B |
| references/cli-reference.md | markdown | 8.2 KB |
| references/threshold-tables.md | markdown | 7.9 KB |
references/threshold-tables.md
View on GitHub# Arista EOS Threshold Tables
Detailed threshold definitions for EOS device health parameters. EOS runs on
Linux — memory thresholds must account for page cache behavior. MLAG and
VXLAN/EVPN thresholds apply only to data center deployments. Adjust thresholds
per device role and traffic profile.
## CPU Thresholds
### System CPU 5-Minute Average
| Level | Range | Meaning | Action |
|-------|-------|---------|--------|
| Normal | 0–40% | Typical operating range | No action |
| Warning | 41–70% | Elevated load | Identify top agent via `show processes top once` |
| Critical | 71–90% | High load, protocol timer risk | Investigate top agents, reduce load |
| Emergency | > 90% | Control plane at risk | Immediate triage, collect `show tech-support` |
### Per-Agent CPU
| Level | Range | Meaning | Action |
|-------|-------|---------|--------|
| Normal | < 30% | Healthy agent | No action |
| Warning | 30–50% | Agent under load | Monitor, identify workload source |
| Critical | 50–80% | Agent overloaded | Investigate protocol/feature causing load |
| Emergency | > 80% | Agent at risk of watchdog timeout | Immediate investigation, may crash |
## Memory Thresholds
### System Memory (Free %)
EOS uses Linux page cache aggressively. "Used" memory includes reclaimable cache.
Use `bash free -m` `available` column for true available memory.
| Level | Free % (show version) | Available % (bash free) | Action |
|-------|----------------------|------------------------|--------|
| Normal | > 40% | > 50% | No action |
| Warning | 20–40% | 30–50% | Monitor trend, identify large consumers |
| Critical | 10–20% | 15–30% | Investigate growth, check for agent leaks |
| Emergency | < 10% | < 15% | Risk of OOM kills, plan reload, open TAC case |
### Flash Storage
| Level | Used % | Meaning | Action |
|-------|--------|---------|--------|
| Normal | 0–70% | Healthy | No action |
| Warning | 71–85% | Filling | Clean old images, archives |
| Critical | 86–95% | Near full | Remove unused files, compress logs |
| Emergency | > 95% | Full — upgrades and configs at risk | Immediate cleanup |
## Agent Health Thresholds
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | All agents running, uptimes match device | Healthy | No action |
| Warning | Agent restarted once (uptime < device uptime) | Recovered from crash | Check agent logs for cause |
| Critical | Agent in crashed/not running state | Subsystem down | Collect logs, attempt restart, escalate |
| Emergency | Multiple agents crashed or restart loop | Systemic instability | Collect show tech, open TAC case |
### Agent Restart Frequency
| Level | Restarts/24hr | Meaning | Action |
|-------|---------------|---------|--------|
| Normal | 0 | Stable | No action |
| Warning | 1 | Single incident | Review logs, monitor |
| Critical | 2–3 | Recurring issue | Investigate root cause, consider image upgrade |
| Emergency | > 3 | Persistent instability | Collect diagnostics, open TAC case |
## Interface Thresholds
### Error Rate
| Level | Rate | Meaning | Action |
|-------|------|---------|--------|
| Normal | < 0.001% | Clean | No action |
| Warning | 0.001–0.01% | Low error rate | Monitor, check optics DOM |
| Critical | 0.01–0.1% | Significant errors | Investigate L1: cable, SFP, fiber |
| Emergency | > 0.1% | High error rate | Replace cable/optic, check remote end |
### Output Discards Per Hour
| Level | Range | Meaning | Action |
|-------|-------|---------|--------|
| Normal | 0–50/hr | Minimal | No action |
| Warning | 51–500/hr | Moderate congestion | Review QoS, check bandwidth |
| Critical | 501–5,000/hr | Significant drops | Tune QoS, consider link upgrade |
| Emergency | > 5,000/hr | Severe congestion | Immediate QoS intervention or upgrade |
### CRC Errors Per Hour
| Level | Range | Meaning | Action |
|-------|-------|---------|--------|
| Normal | 0 | Clean signal | No action |
| Warning | 1–10/hr | Intermittent L1 issue | Monitor, check SFP DOM readings |
| Critical | 11–100/hr | Active L1 problem | Replace cable, clean fiber, reseat SFP |
| Emergency | > 100/hr | Severe L1 failure | Replace cable and SFP, check patch panel |
## MLAG Thresholds
### MLAG Overall State
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | Active, connected, peer-link up, consistent | Fully healthy | No action |
| Warning | Config-sanity inconsistency | Mismatch risk | Resolve config differences |
| Critical | Peer-link down or negotiation disconnected | MLAG degraded | Investigate physical link, check heartbeat |
| Emergency | Split-brain (both active, disconnected) | Traffic duplication/loops possible | Emergency: restore peer-link connectivity |
### MLAG Interface State
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | active-full | Both peers have member up | No action |
| Warning | active-partial | One peer's member is down | Investigate down member |
| Critical | disabled | MLAG interface down | Check physical interface, port-channel |
| Emergency | Multiple interfaces in disabled state | Widespread MLAG failure | Investigate common cause |
### MLAG Heartbeat
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | Heartbeats received on schedule | Healthy control plane | No action |
| Warning | Occasional missed heartbeats | Intermittent issue | Check management network |
| Critical | Sustained heartbeat loss | Control plane risk | Verify management VRF, peer IP |
| Emergency | Complete heartbeat loss with peer-link up | Split-brain risk | Immediate investigation |
## VXLAN/EVPN Thresholds
### VTEP Reachability
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | All expected VTEPs present | Full fabric connectivity | No action |
| Warning | 1–2 VTEPs missing | Partial fabric degradation | Check missing VTEP underlay and EVPN |
| Critical | > 10% of VTEPs missing | Significant fabric issue | Check underlay IGP, spine connectivity |
| Emergency | Vxlan1 interface down | Overlay completely down | Check source interface, underlay routing |
### BGP EVPN Peer Health
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | All peers Established, stable route count | Healthy control plane | No action |
| Warning | Route count deviation > 10% | Route churn | Investigate upstream changes |
| Critical | Peer not Established | EVPN peering broken | Check transport, ASN, route-map |
| Emergency | Multiple peers down | Fabric control plane failure | Check spine, underlay, BGP config |
## Environment Thresholds
### Temperature
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | Below warning threshold | Within operating spec | No action |
| Warning | Above warning, below critical | Approaching limit | Check fans, airflow, ambient cooling |
| Critical | Above critical, below shutdown | Thermal stress | Fix cooling immediately |
| Emergency | At shutdown threshold | Imminent shutdown | Emergency cooling, power off non-essential |
### Power Supply
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | All PSUs OK, redundancy active | Fully protected | No action |
| Warning | Single PSU, no redundancy | Unprotected | Plan PSU installation |
| Critical | Redundant PSU failed | Lost redundancy | RMA failed PSU |
| Emergency | Single PSU degraded | Power loss risk | Emergency maintenance |
### Fan/Cooling
| Level | Condition | Meaning | Action |
|-------|-----------|---------|--------|
| Normal | All fans OK, airflow correct | Adequate cooling | No action |
| Warning | Fan speed elevated or airflow mismatch in rack | Reduced efficiency | Check ambient, correct airflow |
| Critical | One fan failed | Thermal risk | RMA fan, monitor temperature |
| Emergency | Multiple fans failed | Thermal shutdown risk | Reduce load, emergency service |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |