MCP-сервер Uptimepage

• Имя в реестреdev.uptimepage/uptimepageподтверждён в реестре MCP

Имя в реестре лежит в пространстве uptimepage.dev. Реестр пускает в такое пространство только того, кто доказал владение доменом через DNS или HTTP.

• Состояниеотвечаетпроверено 2026-10-19
• Инструменты39
• Сервер называет себяuptimepage
• Версия сервера1.4.0
• Версия протокола2025-06-18

2026-10-19 сервер ответил на initialize и tools/list без входа. Первый ответ за 747 мс.

• Отвечалидёт замерпроверок пока 1, долю покажем после 7
• Медиана первого ответа731 мс
• Место в контексте агента≈ 16 134 токеновописания всех инструментов (39), агент держит их в контексте на каждом ходу
• Описания инструментовA · 99 из 100

• Параметры без описания: в 3 из 39

Программа проверяет сервер раз в день: initialize и tools/list, без модели. Токены - это длина описаний из расчёта примерно четыре символа на токен. Оценка - среднее шести проверок имени, описания, схемы и аннотаций каждого инструмента. Она говорит, насколько понятно описаны инструменты, а не о том, работают ли они.

• acknowledge_incident

  Acknowledge an incident: take ownership and halt escalation. Internal/operational only — does NOT post anything to the public status page. Use post_incident_update for customer-facing updates. Asks for confirmation where the client can…

• add_status_page_components

  Add monitors to a status page as public components, in one confirmation where the client can show a prompt. Give each a public_name the page's readers will understand, since the monitor's own name is operator-facing, and a public_group to…

• cancel_maintenance

  Cancel a maintenance window that has not ended. It leaves the status pages, stops holding paging, and stays listed under past as a record; subscribers already told about it are not sent a cancellation. To stop a running window early and…

• create_maintenance

  Schedule a maintenance window for planned work on some monitors. While it runs they show as under maintenance on every published status page that carries them, and unless suppress_alerts is false nobody is paged for them: checks keep…

• create_monitor

  Create a monitor for an http, tcp, ping, dns, tls_cert, domain_expiry, heartbeat or manual check. A manual monitor has nothing to probe; set its state afterwards with set_monitor_state. The check is run once before anything is saved and…

• create_monitors

  Create several monitors at once, with one confirmation covering the batch where the client can show a prompt. Every check is run once first and all the results are shown together, so a misconfigured endpoint is visible before anything is…

• create_status_page

  Create a status page. It is created unpublished unless you pass enabled, so its components can be curated before anyone can read it. The slug is the page's public address: it is first-come across the platform and moving it later breaks…

• get_flow_runs

  A browser flow monitor's recent runs over a window (1h/24h/7d/30d): every declared step with its outcome and duration, the step a failure stopped on, and the page the browser saw. Use this to answer why a login check failed. Read-only.

• get_flow_step_trend

  How long each step of a browser flow monitor takes over a window (1h/24h/7d/30d), and how far it has moved: per step the earliest and latest mean duration, their ratio, and how many runs passed or failed it. Use this to spot a step…

• get_incident

  One incident: affected monitor, severity, open/resolved times, error sample, the full operator-update timeline, and the status pages an incident with no monitor is posted to. Read-only.

• get_incident_metrics

  Incident metrics over a trailing window (default 30 days): MTTA/MTTR in seconds, total incidents, counts by severity and state, auto- vs human-resolved, and the noisiest monitors. Read-only.

• get_maintenance

  One maintenance window: title, description, start and end, phase, the monitors it covers, and whether it holds their paging. Read-only.

• get_monitor

  One monitor's full configuration — everything the check asserts (expected status, body match, headers, timeout, redirect and TLS policy), the regions it probes from, and how it alerts (failing checks before it pages, whether recovery is…

• get_monitor_history

  One monitor's history over a window (1h/24h/7d/30d): uptime, latency series, a per-region split of the same window, failures with error text, and incident windows. Pass `region` to narrow it to one probe region and tell a partial outage…

• get_org_health

  Org health summary: per-state monitor totals and the worst currently-failing monitors. The one-shot answer to 'what is broken right now?'. Read-only.

• get_org_usage

  Which org this connector is bound to, and the account's resource usage against plan limits: monitors, status pages, members, components, and key policy values. Caps are pooled across every org the account owns, so the counts can exceed…

• get_status_page

  One status page: name, public URL, enabled, and its components with each linked monitor's current state. Read-only.

• list_incidents

  List the org's incidents: incident id, affected monitor, severity, open/resolved times, and latest update phase. Defaults to currently-open ones; pass state="all" with an optional from/to window (default: last 30 days) for resolved…

• list_maintenance

  List maintenance windows: title, start and end, phase (upcoming, active, completed or cancelled), the monitors each covers, and whether it holds their paging. status narrows to active, upcoming or past (ended or cancelled); the default is…

• list_monitors

  List monitors with optional state/type/tag filters and cursor pagination. Each item carries its current state and last-checked time. Read-only.

• list_notification_channels

  The org's notification channels: id, operator-set name, kind (email, slack, telegram, webhook, and so on), and whether the channel is enabled. Two flags say a channel is not working even where it reads as ready: awaiting_verification for…

• list_regions

  The fleet's probe regions: id, display name, city, country, continent, and whether each is on by default for a new monitor. Reports `max_regions` only when the plan reaches fewer regions than the catalog lists, so a set too large to be…

• list_status_page_viewers

  List who signed in to a private status page: each viewer's address, first and last visit, saved passkeys, whether they subscribe by email, and whether they are blocked; and for the page password, which names nobody, the devices signed in…

• list_status_pages

  List the org's status pages: slug, name, public URL, enabled. Cursor-paginated. Read-only.

• list_tags

  Every tag in use across the org's monitors, most-used first, with how many monitors carry each. Pass one back as the `tag` filter to list_monitors. Read-only.

• list_variables

  The org's reusable variables: key, and whether it is a secret. Values are never returned, and a secret's value is never even read. Write a variable into a monitor's header or body as `{{ key }}`, which is resolved when the check runs…

• pause_monitor

  Pause a monitor (stop its checks until resumed). For planned work, create_maintenance holds paging instead while the checks keep running. Asks for confirmation where the client can show a prompt; otherwise runs on the token's scope. Not…

• post_incident_update

  Post a public, customer-facing update to an incident's status-page timeline (phase + message). This is what your subscribers and status-page visitors see. Asks for confirmation where the client can show a prompt; otherwise runs on the…

• post_maintenance_update

  Post an update to a maintenance window: say the work is running late, that it finished, or what changed. Works on an upcoming, running or completed window (a completed window is read-only otherwise, so this is how to add a closing note)…

• publish_incident

  Publish an incident so it appears on every status page carrying the affected monitor, optionally seeding the public title and description. An incident with no monitor needs `status_page_ids` naming the pages to show it on; the list…

• resolve_incident

  Resolve an incident (mark the operational state resolved). Internal only — does not post to the public status page. Asks for confirmation where the client can show a prompt; otherwise runs on the token's scope. Not read-only.

• resume_monitor

  Resume a paused monitor (restart its checks). Asks for confirmation where the client can show a prompt; otherwise runs on the token's scope. Not read-only; idempotent.

• run_check_now

  Run a check on a monitor immediately and record the result. Asks for confirmation where the client can show a prompt; otherwise runs on the token's scope. A down result may fire the org's normal alerts. Heartbeat monitors cannot be probed…

• set_monitor_state

  Set the state of a manual monitor: up, degraded or down. A manual monitor has no probe; its state is whatever an operator last set, for a service only people can judge, such as a SIP trunk, a carrier link or a partner's back office. Down…

• unpublish_incident

  Hide a published incident from the public status pages again. Its operator timeline is untouched. Asks for confirmation where the client can show a prompt; otherwise runs on the token's scope. Not read-only; idempotent.

• update_maintenance

  Edit an upcoming or running maintenance window: its title, description, start, end, the monitors it covers (monitor_ids replaces the whole set) or whether it holds their paging. end_now ends a running window at once and keeps it as…

• update_monitor

  Change how loudly a monitor is watched: check interval, alert confirmations, recovery notices, reminder interval, tags, group, the multi-region detection quorum, and which notification channels it alerts (channel_ids replaces the whole…

• update_status_page

  Rename a status page, move it to a new slug, publish and unpublish it, or change who may open it. An omitted field is left alone. Changing the slug moves the public URL and breaks existing links. access replaces the page's access settings…

• update_status_page_component

  Change how one monitor is presented on a status page: its public name, description, group or position. An omitted field is left alone. Not read-only; idempotent.