Changelog

What's new in Observe, newest first. Atom feed

November 2026

A routine renewal from the same certificate authority is now recorded on the monitor's SSL timeline without an alert, so automatic Let's Encrypt and Cloudflare renewals stay quiet. You are still alerted when the certificate authority changes, the key, signature or list of names gets weaker, or the new certificate expires sooner than the old one or within your expiring-soon warning. To be alerted on every change as before, turn on Alert on every certificate change in the monitor's SSL settings (ssl_alert_on_every_change in the API and MCP).

Organization → Billing shows your plan, price, next renewal (or the date it ends) and how many monitors you use of your limit, with Manage billing and Change plan. After you subscribe, Stripe brings you back to this page and it shows your new plan as soon as it is active, instead of the monitors list.

Every uptime check now runs from US East, EU West and US West at once, and a monitor only counts as down when two of the three agree, so one bad network path no longer pages you. Each monitor shows how each region saw it on its last check.

Add Observe as a connector with just its URL, https://observe.live/mcp: you sign in, pick the organization, read only or read and write, and optionally which monitors, then click Allow. No token to copy. Connected apps are listed in Profile → API & data next to your API tokens, and Disconnect stops one at once. API tokens keep working on the MCP server as before. Every MCP tool now has a title and says whether it only reads, changes or deletes data, or reaches outside Observe, so clients can ask before the risky ones.

Jira Service Management Operations, Opsgenie's successor, is a new alert channel: alerts open, get still-down notes, are acknowledged when you acknowledge in Observe, and close on recovery. Opsgenie and JSM channels now set a priority (P1 for critical monitors, or one you choose), an optional responder team, the monitor as entity, tags and details. Opsgenie reaches end of support on April 5, 2027: each Opsgenie channel has a one-click Switch to JSM, and the migration notes are in the docs.

Add a passkey in Profile → Security and sign in with Touch ID, Face ID, Windows Hello, your phone or a security key, no password needed. A passkey sign-in counts as two-factor, including for organizations that require it, and two-factor users can answer the code prompt with a passkey instead.

The app header is now a single row, with the breadcrumb above the page title and a slimmer footer. Workspace navigation moved into a top bar, and the organization switcher shows each organization's icon.

Every monitor section opens with one summary card in the style of Uptime: title, subtitle and actions in its header. Section tabs have icons, and the monitor header has real buttons for Snooze alerts, Edit and Remove.

Every notification channel has a Send test button, so you can confirm Slack, email or a webhook works before a real outage.

Scheduled uptime checks now run in batches, so every monitor is checked on time even on the busiest minute. A check interrupted mid-run is retried once instead of being dropped.

Public status pages are served from a cache close to your visitors, so they stay fast when everyone is refreshing during an outage. Subscribing sends at most three confirmation emails to an address a day.

We are testing uptime checks that run from three regions in North America and Europe and only alert when two of them agree a site is down. Until it ships, checks run from one location and re-check before any alert.

Each channel can send a kind of alert (uptime, SSL, domains and DNS, servers, health checks, content) right away, in a daily digest, or not at all. Set quiet hours in your organization's timezone; down alerts for monitors marked critical, PagerDuty and Opsgenie always come through.

Alert emails can carry your organization's logo, accent colour and footer text, or use a plain-text design.

Scripts written for the Oh Dear API can list and create monitors, request a check run, snooze a check and read uptime and downtime from Observe.

Watch each uptime check go from queued to running to finished, with late checks flagged.

Owners can hand an organization to another member. The only owner can't delete their account while others still use the organization, and deleting an account removes the organizations nobody else is in.

Incident alerts, resolved emails, status page incident posts and maintenance notices give their times in your organization's timezone.

Signing up from an invitation starts with the invited email and confirms you joined the team. Your new plan is active as soon as you return from checkout, and a pending cancellation shows the date the plan ends.

Improved

A new look

The site and the app have a fresh design, and your light or dark theme choice now sticks as you move between pages.

If a site stays down, Observe reminds you after an hour, after four hours, then once a day. Paused monitors stay quiet, and failed scheduled tasks no longer show on public status pages.

Laravel apps using spatie/laravel-schedule-monitor can sync their scheduled tasks into Observe as heartbeat monitors.

Bring your monitors over from Oh Dear, alongside the UptimeRobot, Better Stack and CSV imports.

Leave notes for your team on an incident's timeline; they never appear on a status page.

The API is described by an OpenAPI 3.1 spec at /api/openapi.json, and the app shows where to get a token for the MCP server.

Another admin can take over an organization's billing without contacting support.

Application health checks and server SSL send a recovery alert when they pass again. Each server gets its own PagerDuty or Opsgenie incident, which closes on its own when the server recovers.

Server ping falls back to a TCP connect where ICMP is blocked, instead of reporting the server down. A server with no results yet shows as Pending, and a heartbeat that hasn't pinged yet shows as waiting.

Monitor pages load in about half a second, down from up to two.

DNS record checks look up DMARC and DKIM records where they actually live.

Each recipient gets one email per alert, and heartbeats and checks no longer leave silent gaps. Business-hours channels don't get a recovery for an outage they never heard about, and ending a snooze early still alerts for an outage that began during it.

Members can leave an organization themselves, and status page subscribers can unsubscribe in one click from their mail app. Uptime intervals go up to 60 minutes, and page headers and status page history fit on a phone.

Fixed a security issue in how checks make outbound requests.

Follow what changes here or in the Atom feed. Owners and admins also get a monthly what's-new email they can unsubscribe from in one click.

The pricing page answers what the trial includes, how a re-check works before an alert, and annual billing.

A guide at /docs covers alert channels, status pages and cron heartbeats.

A short wizard walks new accounts from adding a site to setting up alerts and seeing the first check.

A failed card renewal keeps your plan while it is retried, with a banner to update the card, and downgrading pauses extra monitors instead of deleting them. Checkout accepts promotion codes, and the billing and refunds policy is at /billing-policy.

Improved

Branded emails

Every email from Observe has a branded header and a proper footer.

Side-by-side comparisons with Oh Dear, UptimeRobot and Better Stack.

The REST API now covers incidents, check runs, snoozes, every check type's results, maintenance windows, status pages and servers, with public docs at /docs/api. An MCP server lets AI assistants read and act on your monitors with the same permissions as your API token.

Choose Free, Pro or Business, with a 14-day trial that needs no card. Owners get an email three days before the trial ends and when it ends.

Organizations can sign in through their identity provider with SAML, and can require it for every member.

Run several status pages per organization, in English, Spanish, French, German, Dutch or Portuguese. Pages can be private behind a secret link, post incidents automatically, and publish a JSON feed.

Tag monitors and route alerts by tag group, so each team hears about its own sites.

Alerts are retried when a channel fails, failures are tracked, a broken channel is paused automatically, and webhooks are signed.

Uptime checks take a timeout, redirect rules and an alert delay in minutes; scheduled tasks accept starting and failed pings with a run history; application health can be pulled. DNS checks query every authoritative nameserver and alert when they disagree, SSL checks flag certificates browsers reject, blocklists include IP lists, and Lighthouse runs on desktop or mobile with per-metric thresholds. Crawls take URL ignore patterns and an external-link toggle, sitemap checks follow every index, and servers can opt in to a daily full port scan.

The product has a new name, Observe, and a new homepage. Reaching your plan's site limit shows one clear banner with an upgrade link.

June 2026

A notifications bell and feed show recent alerts inside the app, with unread ones marked.

SSL, DNS, domain, blocklist, performance and heartbeat pages explain their results in plain language and compare them with earlier runs.

Send alerts to Discord, Telegram, Microsoft Teams, Pushover and ntfy. PagerDuty and Opsgenie incidents open when a monitor goes down and close when it recovers.

Uptime checks can assert that text is absent, check JSON values and response headers, and show a timing breakdown for each response. Heartbeats can follow a cron schedule in your timezone, and ping, TCP port, domain and application health checks report more detail.

Post status updates tagged by severity, show uptime history bars, schedule maintenance windows, and let visitors subscribe for updates.

Version 1 of the REST API lets you manage monitors from your own scripts.