skills/automateyournetwork/netclaw/te-network-monitoring

te-network-monitoring

SKILL.md

ThousandEyes Network Monitoring

Query Cisco ThousandEyes network intelligence via two MCP servers — the community server (9 tools, local stdio) for core monitoring and the official server (~20 tools, remote HTTP) for advanced analysis. Monitor synthetic tests, inspect agent health, view path visualizations, query dashboards, and investigate network events.

MCP Servers

Community Server (local, stdio)

  • Repository: CiscoDevNet/thousandeyes-mcp-community
  • Transport: stdio (Python)
  • Install: git clone + pip install -r requirements.txt
  • Script: src/server.py
  • Python: 3.12+
  • Requires: TE_TOKEN
  • Status: Alpha/MVP, read-only

Official Server (remote, HTTP)

  • Repository: CiscoDevNet/ThousandEyes-MCP-Server-official
  • Endpoint: https://api.thousandeyes.com/mcp
  • Transport: Remote HTTP (via npx mcp-remote)
  • Auth: Bearer token (Authorization: Bearer <TE_TOKEN>)
  • No local install — hosted by Cisco
  • Requires: TE_TOKEN, Node.js (for npx), org not opted out of AI features

Community Server Tools (9)

Tool Parameters What It Does
te_list_tests aid?, name_contains?, test_type? List all configured tests with filtering by name, type, or account group
te_list_agents agent_types?, aid? List enterprise, enterprise-cluster, and cloud agents
te_get_test_results test_id, test_type, window?/start?/end?/aid?/agent_id? Fetch network, page-load, and web-transaction test results
te_get_path_vis test_id, window?/start?/end?/aid?/agent_id?/direction? Path visualization — hop-by-hop network path data
te_list_dashboards aid?, title_contains? List all ThousandEyes dashboards
te_get_dashboard dashboard_id, aid? Dashboard details including widget inventory
te_get_dashboard_widget dashboard_id, widget_id, window?/start?/end?/aid? Widget data for a specific dashboard widget
te_get_users none List users in the ThousandEyes account
te_get_account_groups none List account groups accessible to the authenticated org

Official Server Tools (~20)

Core Monitoring

Tool What It Does
List Tests View all configured tests (web, network, DNS, voice)
Get Test Details Detailed information about a specific test
List Events Find network and application problems within time ranges
Get Event Details Deep dive into specific events with impacted targets
List Alerts View triggered or cleared alert rules
Get Alert Details Comprehensive alert information with conditions and targets
Search Outages Find network and application outages with filters
Instant Tests Active troubleshooting — run tests on demand

Advanced Analysis

Tool What It Does
Get Anomalies Detect metric anomalies in test data over time
Get Metrics Retrieve aggregated metrics for dashboards and reports
Views Explanations AI-powered — explain specific test results and visualizations

Endpoint Monitoring

Tool What It Does
List Endpoint Agents and Tests List endpoint agents with filtering
Get Endpoint Agent Metrics Time series data from endpoint agents

Network Path & BGP Analysis

Tool What It Does
Get Path Visualization Network paths and hop-by-hop routing
Get Full Path Visualization Comprehensive path data for all agents
Get BGP Test Results BGP reachability and routing information
Get BGP Route Details Detailed AS path and routing information

Account Management

Tool What It Does
Get Account Groups List available account groups

Key Concepts

Concept What It Means
Test Synthetic monitoring probe — HTTP, page load, web transaction, network (ICMP/TCP), DNS, voice, SIP
Agent ThousandEyes vantage point — Enterprise (on-prem), Enterprise Cluster (distributed), Cloud (Cisco-hosted global)
Path Visualization Hop-by-hop network path from agent to target — shows each router, latency per hop, loss, MPLS labels
Event Detected network or application problem — outage, latency spike, packet loss, DNS failure
Alert Rule Threshold-based trigger — fires when metrics exceed defined conditions
Dashboard Custom visualization with widgets — aggregated views of test data, metrics, and alerts
Endpoint Agent Lightweight agent on user workstations — monitors real user experience (WiFi, VPN, SaaS)
Instant Test On-demand test execution — active troubleshooting without waiting for scheduled test cycles
BGP Route BGP reachability data — AS path, prefix, origin, route visibility from ThousandEyes BGP monitors

Workflow: Network Performance Assessment

When a user asks "how's the network performing?":

  1. List tests: te_list_tests — discover all configured monitoring tests
  2. Test results: te_get_test_results for key network tests — latency, loss, jitter
  3. Alerts: List Alerts (official) — any active alert conditions
  4. Events: List Events (official) — recent network/application problems
  5. Dashboards: te_list_dashboards + te_get_dashboard_widget — executive summary metrics
  6. Report: network performance dashboard with test-by-test metrics and alert status

Workflow: Path Troubleshooting

When investigating "traffic to example.com is slow from the London office":

  1. Find test: te_list_tests filtered by name or target
  2. Path visualization: te_get_path_vis for the relevant test — hop-by-hop path
  3. Full path: Get Full Path Visualization (official) — all agents, all paths
  4. BGP routes: Get BGP Route Details (official) — is traffic taking an unexpected AS path?
  5. Anomalies: Get Anomalies (official) — metric deviations over time
  6. Report: path analysis with hop-by-hop latency, loss source, and routing assessment

Workflow: Outage Investigation

When responding to a ThousandEyes alert:

  1. Alerts: List Alerts (official) — active alert details, affected tests
  2. Events: List Events (official) + Get Event Details — timeline, impacted targets
  3. Outages: Search Outages (official) — broader outage scope (ISP, CDN, SaaS)
  4. Path vis: te_get_path_vis — where in the path is the problem?
  5. Instant test: Instant Tests (official) — run on-demand tests from multiple agents
  6. Report: outage timeline with root cause (ISP node X, hop Y, BGP route change)

Workflow: Endpoint Experience

When investigating "users complain about slow VPN":

  1. Endpoint agents: List Endpoint Agents and Tests (official) — find agents on affected users
  2. Metrics: Get Endpoint Agent Metrics (official) — WiFi signal, VPN latency, DNS response
  3. Path vis: Get Path Visualization (official) — user-to-VPN-gateway path
  4. Correlate: compare with enterprise agent test results — is it user-side or network-side?
  5. Report: endpoint experience analysis with WiFi, VPN, and path diagnostics

Workflow: BGP Monitoring

When auditing BGP health:

  1. BGP tests: te_list_tests filtered by type=bgp
  2. BGP results: Get BGP Test Results (official) — reachability, prefix visibility
  3. Route details: Get BGP Route Details (official) — AS paths, origin validation
  4. Anomalies: Get Anomalies (official) — BGP metric deviations
  5. Report: BGP health summary with route stability assessment

Integration with Other Skills

Skill How They Work Together
pyats-routing ThousandEyes external BGP view + pyATS internal routing state = full picture
pyats-troubleshoot ThousandEyes path vis (internet view) + pyATS show commands (device view)
aws-network-ops ThousandEyes cloud agent tests + AWS VPC flow logs for hybrid visibility
gcp-cloud-monitoring ThousandEyes metrics + GCP Cloud Monitoring for cross-platform correlation
meraki-monitoring ThousandEyes synthetic tests + Meraki Dashboard live diagnostics
fmc-firewall-ops ThousandEyes path vis through firewall + FMC rule audit
servicenow-change-workflow ThousandEyes alerts trigger ServiceNow incidents
gait-session-tracking Record all ThousandEyes investigations in GAIT audit trail
slack-network-alerts Deliver ThousandEyes alert summaries to Slack channels

Important Rules

  • Read-only — community server is read-only; official server has Instant Tests (on-demand test execution)
  • API rate limits — ThousandEyes API usage counts against your org's rate limit; avoid rapid-fire queries
  • Account group scope — queries default to the authenticated user's account group; specify aid for cross-group queries
  • AI features — official server requires org not opted out of ThousandEyes AI features
  • Time windows — use window (seconds) or start/end (ISO 8601) for time-bounded queries
  • Record in GAIT — log all ThousandEyes investigations for audit trail

Environment Variables

  • TE_TOKEN — ThousandEyes API v7 OAuth bearer token (used by both community and official servers)
Weekly Installs
1
GitHub Stars
282
First Seen
10 days ago
Installed on
mcpjam1
claude-code1
replit1
junie1
windsurf1
zencoder1