Terminal
Terminal/Skill
Generate a one-shot health snapshot — monitor counts by state, which are currently down, recent incident count, and total downtime — useful as a daily brief or ad-hoc check.
By UptimeRobot
Works with
npx skills add uptimerobot/ai --skill monitor-health-summaryPreflight — read first. If you cannot see any
uptimerobot:*MCP tools in your tool list, invoke theuptimerobot:setupskill before doing anything else. Do not tell the user the MCP is misconfigured —setup's Step 0 detects the common case (server connected, tools loaded after session start) and resolves it without re-keying.
A fast overview of your monitoring account's current state. Combines account-wide stats, live down/paused lists, and recent incident activity into a single report.
Use this when the user asks "how's everything looking?", "give me a health check", "what's our monitoring status?", or "morning standup brief".
{ "timeRange": "24h" }
Call get-monitor-stats. Returns:
up, down, paused, notStarted).Adjust timeRange if the user wants a longer window (e.g. "7d" for a weekly brief).
{ "filter": ["DOWN"], "limit": 50 }
Paginate with cursor until hasMore: false. Collect each monitor's id and friendlyName and the timestamp from get-monitor-details if you need exact downtime duration.
If there are zero DOWN monitors, skip step 3 and say so clearly.
{ "filter": ["PAUSED"], "limit": 50 }
Paginate until hasMore: false. You only need names and count — no need to fetch details for each.
{ "timeRange": "24h", "limit": 25 }
Call list-incidents. Paginate until hasMore: false. Collect total count and sum of incident durations for the report footer.
For each DOWN monitor (limit to the top 3–5 by name if there are many), check response times to distinguish a full outage from degradation:
{ "monitorId": 800123456, "timeRange": "1h", "bucketSize": 60 }
Only call this when the user wants more detail or when the monitor count is small. Avoid looping over dozens of monitors.
Health summary — 2026-05-01 (last 24h)
Status: 2 DOWN · 3 PAUSED · 42 UP (47 total)
Uptime: 98.7% across active monitors
Incidents last 24h: 4 (total downtime: 2h 13min)
Currently DOWN (2):
• prod-api (800123456) — down since 14:02 UTC (~58 min)
• staging-db (800123457) — down since 13:45 UTC (~1h 15min)
Paused (3): deploy-preview, load-test-env, sandbox-api
If everything is up:
Health summary — 2026-05-01 (last 24h)
Status: 0 DOWN · 1 PAUSED · 46 UP (47 total)
Uptime: 100.00% across active monitors
Incidents last 24h: 0
All active monitors are up. One monitor paused: sandbox-api.
NOT_STARTED monitors as "down" — they're unstarted, not failing. Exclude them from down/up tallies or list them separately.PAUSED monitors as "healthy" — they're unmonitored. Always surface them so the user is aware.get-response-times over every UP monitor — expensive and unnecessary. Reserve it for DOWN or flagged monitors.list-monitors or list-incidents — with many monitors or incidents you'll silently under-count.notStarted with paused — they're different states. notStarted means the monitor was never activated.incident-response — to triage what the summary surfaces as down.stats — simpler alternative when the user only wants uptime % without the full breakdown.incidents — to list and investigate individual downtime events.bulk-pause — to act on the paused-monitor list (resume all at once).conventions — time range formats and pagination patterns.