Watchgoose vs Uptime Kuma: hosted or self-hosted monitoring?
The first decision is not feature count but ownership. Uptime Kuma is MIT-licensed software that you run on your infrastructure. It monitors websites, TCP services, DNS, databases, containers, Push heartbeats, and more, with status pages on the same self-hosted deployment. Watchgoose is a commercial hosted service focused on scheduled work that can send a signal; the subscription covers operating that monitoring service.
As a Uptime Kuma alternative for cron jobs, Watchgoose exchanges infrastructure control for Cron and systemd OnCalendar expectations, paid-plan team access, and managed check and history limits. Uptime Kuma Push monitors use a rolling interval, while Watchgoose can evaluate a simple interval, cron expression, or OnCalendar expression plus grace. Both can receive explicit failure signals. The practical question is whether broad self-hosted uptime monitoring is worth responsibility for deployment, upgrades, backups, capacity, availability, and the alert path itself.
Competitor facts in this introduction: README, licence, Push API docs.
Start monitoring
Start with 10 checks on the permanent Free plan.
Comparison
| Question | Watchgoose | Uptime Kuma |
|---|---|---|
| Delivery model | Commercial hosted scheduled-job monitoring. [1] | MIT-licensed, self-hosted monitoring software. Version 2.5.3 was the latest stable release reviewed. |
| Pricing model | Free 10 checks/one user; Startup 100 checks and teams at EUR 17/month; Business 1,000 checks at EUR 73/month. [1] | No software licence fee under MIT. The operator supplies infrastructure and operations. No first-party managed-hosting price was found. |
| Open source program | Qualifying Open Source projects and nonprofits may receive a needs-matched paid plan free. Watchgoose pledges 5% of profit to Open Source projects. [1] | Not applicable: Uptime Kuma is itself MIT-licensed open-source software. |
| Monitor limits | Plan limits are 10, 100, and 1,000 checks for Free, Startup, and Business. [1] | No vendor quota applies to the software. A supported capacity ceiling was not published and depends on the deployment. Since version 2.5.0, monitor intervals can extend beyond the previous 24-day limit. |
| Scheduled-job semantics | Simple, Cron, and OnCalendar schedules plus grace. Missing success and /fail alert; /start optionally records runtime and uses the same grace. [1] | Push monitors use a rolling interval from the latest heartbeat. There is no separate expected period and grace; official guidance is to set the interval to expected period plus grace. Push requests can report status=down. |
| Teams and access | Startup and Business have project teams with Manager and Read-only roles as well as owner/member access. [1][2] | Stable 2.5.3 has no documented multi-user teams or RBAC. Related work remains future/development work. |
| Sign-in options | Email login links and password login remain available. GitHub sign-up and sign-in use the verified primary email; Watchgoose does not retain the GitHub token, account ID, username, or profile data. [1][2][3] | Self-hosted instances create a local admin username and password, with optional TOTP two-factor authentication. Stable 2.5.3 has no GitHub sign-up or sign-in in its product UI. |
| Billing currency and tax | Public EUR prices exclude VAT; commercial billing is handled through Paddle. [1][2] | Not applicable to downloading the MIT software. Infrastructure and any third-party support or hosting are the operator's separate costs. |
| Data location | Europe hosted, operated & owned; not an EU-residency claim. [1] | Chosen by the operator. Docker normally persists application data at /app/data. |
| API | Project-scoped management API with read-write and restricted read-only keys. [1] | Push, Prometheus, badge, and public status-page endpoints exist. The broader internal API is explicitly unsupported for third-party integrations and can change. |
| MCP and Claude | Publicly documented hosted MCP endpoint with OAuth 2.1 for one approved project. Access starts read-only; write access requires explicit authorization and project permission. Watchgoose is listed in Claude's connector directory, listed in ChatGPT plugins for GPT and Codex, and published in the official MCP Registry. [1][2][3][4] | No first-party customer-facing hosted MCP server or Claude connector-directory listing was found in the reviewed stable 2.5.3 repository, README, or official wiki. An MCP proposal was closed, and a separate MCP request remains open. |
| Self-hosting | No public Watchgoose self-host distribution was offered on reviewed pages. [1] | Yes. Self-hosting is the core model, with Docker, Compose, and non-Docker installation guidance. |
| Status pages | One basic opt-in public status page per project is available on all plans. It shows up to 100 owner-selected checks with ordered client-safe public names, optional groups, current state, and last-result age. [1] | Supports multiple public status pages and domain mapping. Public status results are cached and refreshed on a five-minute cycle. |
| Wider monitoring | Focused on scheduled work that can send a signal. It does not automatically observe arbitrary jobs or replace website, network, or database probes. [1] | Also monitors HTTP(S), TCP, keywords/JSON queries, WebSockets, ping, DNS, Docker, databases, game servers, and more. |
Choose Uptime Kuma if...
- You want fully self-hosted monitoring with no software licence fee and accept responsibility for operating it.
- You want website, network, database, container, and simple push monitoring in one interface.
- A rolling "heartbeat within N seconds" model fits your recurring jobs.
- You need multiple status pages on infrastructure you control.
- Broad notification integrations matter more than stable team roles or a supported management API.
Sources: README, status-page docs, internal API docs, multi-user issue.
Uptime Kuma is not a lesser option because it is free software. It is the better choice for operators who actively want self-hosting and already have a credible plan for upgrades, backups, storage, availability, and alert delivery.
Choose Watchgoose if...
- You want a hosted MCP connection listed in Claude's connector directory, with read-only access by default.
- You want GitHub sign-up and sign-in instead of a local-only account.
- You do not want to operate the monitoring service that should detect failures in your other systems.
- You need Cron or systemd OnCalendar expectations and a separate schedule-plus-grace model.
- You need paid-plan project teams and read-only client or teammate access.
- You want fixed check bundles, EUR public pricing, and Paddle-managed billing.
- Your scope is scheduled-job monitoring and plan-limited run history, not broad website and network monitoring or hosted-on-your-infrastructure status pages.
Watchgoose is monitoring-only. It does not run jobs, inspect backups, or validate restorability, and it does not offer a public self-host edition.
Frequently asked questions
Which service offers hosted MCP access?
Watchgoose documents an OAuth-protected MCP endpoint for one approved project and is listed in Claude's connector directory. No first-party hosted MCP server or Claude directory listing was found in Uptime Kuma's reviewed stable 2.5.3 project surfaces; its recent MCP items remain unshipped proposals.
Is Watchgoose a good Uptime Kuma alternative?
It can be for teams that want hosted cron monitoring, native Cron or OnCalendar schedules, and project roles. Uptime Kuma remains stronger when self-hosting, broad uptime probes, multiple status pages, or domain mapping are requirements.
Is Uptime Kuma free?
The software is MIT licensed. Operators still provide and maintain its infrastructure, storage, backups, upgrades, and alert path.
Can Uptime Kuma monitor cron jobs?
Yes, when the job calls a Push URL. Stable 2.5.3 uses a rolling interval and does not evaluate the job's cron or OnCalendar expression.
Does Uptime Kuma support teams and read-only roles?
Stable 2.5.3 has no documented multi-user teams or RBAC. Watchgoose paid plans include a Read-only project role; Uptime Kuma's related multi-user work was still future or development work when reviewed.
Does Uptime Kuma have a supported REST management API?
It exposes several useful endpoints, but the broader internal API is explicitly unsupported for third-party integrations.
FAQ competitor sources: licence, stable source, open MCP request, Push API docs, grace explanation, multi-user issue.
All competitor facts re-verified on 2026-09-03. Tell us if something is outdated.