Changelog

​

New

• Terraform provider. New larmhq/larm provider for managing Larm with Terraform. Resources: larm_monitor, larm_alert_channel, larm_webhook_subscription, and larm_status_page (manages the full structure — groups, components, monitor links — atomically as a single transaction).

​

CLI

• v0.4.0: larm component-groups and a --group flag on components let you manage component groups from the terminal; larm status-pages show now renders the components tree. Breaking: larm components list is removed in v0.4.0 — use larm status-pages show <id> instead.

​

Disruptions

• “Report disruption” modal on the disruptions list (title, impact, started-at, body, page picker).

​

Status pages

• Open Graph and Twitter card images generated automatically for every public page.
• Visual polish on public pages: centered “Powered by Larm” footer, “Theme” swatch removed from the editor header, open-in-new-tab icon and “View public page” menu next to the status page URL.

​

Monitors

• Region labels simplified to eu-central instead of eu-central-1 across the UI.
• Fast path from the landing-page response-time tool to your first monitor: the URL you tested is prefilled at signup and carried into the quickstart.

​

Alerts

• “Suggest monitors” modal after creating an alert channel lets you attach it to existing monitors in one step.
• Retrying a delivery from the notification log now clears the previous error text and stamps the new attempt number.

​

Auth & accounts

• Sender address standardized on hello@larm.dev for auth, alerts, status-page subscriptions, and invitations.
• TOTP manual-entry secret no longer renders with stray whitespace.

​

Public API

• /api/v1/monitors/:id/response-times now aggregates by region (us-east) rather than per location (us-east-1).

​

Marketing

• May blog posts: Dokploy, Railway, SSL certificate expiry, Hetzner, Fly.io.
• About page reframed as a note from the founder.

​

New

• MCP server. Connect Claude Code, Claude Desktop, Cursor, or any other MCP-compatible harness or service to your Larm organization. 12 tools across monitors, alert channels, and disruptions. OAuth 2.1 with PKCE. Available on all plans, including Free.
• Larm CLI. Open-source CLI: brew install larmhq/tap/larm. Manage monitors, alerts, status pages, components, disruptions, and webhooks from your terminal. Browser OAuth or API key, larm describe for LLM agents, raw larm api for any endpoint, table/JSON/--jq/--fields output. Breaking: larm incidents renamed to larm disruptions in v0.2.0.
• Disruptions. Open a disruption from the dashboard, link it to one or more status pages, mark per-component severity, post timeline updates, resolve to auto-publish recovery to every linked page.
• Public OpenAPI spec at /openapi.yaml covering every v1 endpoint.
• Dashboard home with a live globe of probe activity.
• Command palette. Cmd-K on every authenticated page jumps to monitors, alerts, status pages, settings, billing, and more.
• Two-factor authentication (TOTP) with authenticator-app enrollment and 8 recovery codes.
• Passkey sign-in (WebAuthn) with passkey management in Settings.
• OAuth device flow for CLI and third-party clients, with an organization picker on the verify page.
• Display name on your user profile, surfaced across the dashboard.
• Interactive onboarding. Three steps: add a monitor with live URL verification, attach an alert channel, create a status page.
• Analytics consent banner with Accept / Decline on first sign-in.
• Product-update email preference toggle in Settings.

​

Disruptions

• Monitors that go down auto-create a disruption; affected status-page components flip; recovery auto-resolves the disruption.
• A worst-impact status dot in the dashboard header links to disruptions from every page.

​

Status pages

• Public pages auto-refresh when you publish new content — no manual reload.
• Components organize into collapsible groups; drag to reorder.
• Status page list redesigned with a 30-day uptime sparkline per page.
• RSS and Atom feeds for disruptions, surfaced in the Subscribe panel.
• Color picker for the brand color; logo upload with cancel and remove controls.
• Visual polish on public pages: gray dot for maintenance state, “Last updated” timestamp removed from the public footer.

​

Monitors

• Heartbeat graph on the monitor detail page (expected vs received pings per bucket).
• Reorderable monitor groups on the list (drag-to-group, collapsible).
• HTTP monitor secrets: reference with {{VAR}} in headers or body; “Verify” runs against the live config with your unsaved secret values.
• Auto-disable when a monitor stays down for 48 hours or auto-creates more than 10 disruptions in 24 hours, with email warnings before disabling.
• Uptime bars show downtime duration (e.g. “2m 30s”) instead of failed-check counts.
• Fixed: heartbeat monitors that were created but never pinged now correctly transition to pending and then to down at the first missed deadline.

​

Alerts

• Warning banner and bell-slash badge for monitors with no alert channel attached.
• Weekly digest email to owners listing monitors that have been without an alert channel for 7+ days.
• SSL certificate expiry thresholds shifted to 28 / 14 / 7 / 1 / 0 days (was 30 / 14 / 7 / 1 / 0).

​

Auth & accounts

• Sign-up now rejects emails with consecutive dots (foo..bar@example.com); magic-link send failures surface an error instead of silently claiming success.

​

Billing

• MCP and REST API write access enabled on the Free plan.

​

Public API & webhooks

• New endpoints: GET/POST/PATCH /api/v1/disruptions, full CRUD on /api/v1/status-pages/:id/components, GET /api/v1/me.
• Webhook subscription secrets are encrypted at rest.

​

Mobile

• Dashboard home, monitor list and detail, alert channel pages, status page editor, and Settings are now usable on phones (responsive layouts, sticky horizontal nav on Settings, hamburger nav with disruption-status dot, no email-field auto-focus on touch devices).

​

Docs

• New guide: CLI.
• New guide: MCP — connecting Claude Code, Claude Desktop, and Cursor.
• New API reference sections: Disruptions and Status page components.
• Status pages docs rewritten with a component-states table and email-subscriber plan limits.

​

Marketing

• New landing pages: larm.dev/mcp and larm.dev/cli.
• Blog launched with RSS. April posts: Multi-probe voting, TTFB, Cron jobs.
• Shareable response-time tool result page at larm.dev/tools/response-time/r/:id.
• Status pages page rewritten.