Cronalert

What cronalert does

The cronalert MCP server exposes CronAlert account operations as MCP tools for monitoring and incident management. It belongs in the 📊 Monitoring category and is suited to teams that want an AI client to inspect uptime data or make controlled changes to their monitoring configuration.

Read tools can list monitors, retrieve an individual monitor, inspect check history, find incidents for one monitor, list active incidents across the account, view public status pages, and read updates attached to an incident. Write tools can create, modify, pause, resume, and permanently delete monitors. Agents can also post updates to active incidents and import monitor definitions.

Monitor creation supports HTTP checks, heartbeat checks, and TCP port checks. Heartbeat monitors use an expected interval and an optional grace period, making them suitable for scheduled jobs such as backups. Imports can use data from UptimeRobot, Pingdom, Better Stack, StatusCake, Checkly, Oh Dear, Cronitor, Healthchecks.io, Dead Man’s Snitch, Datadog, CronAlert JSON, or CSV.

How it works

The cronalert MCP server sends requests to the CronAlert API at cronalert.com/api/v1/ using an API key. The key is scoped to one team, so operations are limited to that team’s resources. The server does not store data locally; monitor configuration and the authentication key are transmitted to CronAlert when the relevant API operations run.

A read-only API key permits listing and reading monitors, results, and incidents but rejects create, update, delete, and import operations. A read-and-write key enables those changes. If delete confirmation is enabled for the key, delete_monitor first returns a preview and short-lived confirmation token; the agent must submit that token in a second call to complete deletion.

Setup and configuration

Create a CronAlert account and generate an API key under Settings > API Keys. The local setup requires Node.js 18 or newer and the CRONALERT_API_KEY environment variable. With a compatible client, run the package through npx -y cronalert-mcp and pass the key in the client’s environment configuration.

A remote deployment is available at https://cronalert.com/mcp, which uses Streamable HTTP. Clients connecting to that endpoint must send the key as an Authorization: Bearer ca_your_key header. Claude Desktop, Claude Code, Cursor, and Windsurf setup examples are provided by the project. An optional cronalert-uptime-monitoring skill adds guidance for monitor selection, heartbeat grace periods, outage triage, and confirmation before deletion.

Tools and capabilities

Available operations include:

• List monitors with status and response times.
• Create HTTP, heartbeat, or TCP monitors.
• Read or update monitor settings, including heartbeat timing and pause state.
• Delete monitors, subject to the configured permissions and confirmation behavior.
• Retrieve check history, uptime percentages, and response times.
• Review monitor-specific or account-wide active incidents.
• List public status pages and manage updates on active incidents.
• Import monitor definitions from supported services or CSV.

For heartbeat monitors, checkInterval represents the expected number of seconds between pings, while gracePeriod allows additional time before an incident is raised. HTTP and TCP monitor intervals are determined by the CronAlert plan.

Limitations and notes

A read-only key cannot perform write operations regardless of client-side approval prompts. Client approval is therefore not the only permission boundary. Deletion can be permanently destructive, and confirmation-token protection applies only when enabled on a read-and-write key.

The server requires access to the CronAlert API and an account API key. CronAlert processes transmitted data on its infrastructure, including Cloudflare-hosted services, according to its privacy policy.