Every release,
in the open.

• FixRecovery alerts now show the actual duration of the incident that just resolved. Previously a brief incident could report a downtime spanning back to an unrelated earlier outage (for example "54 days").

• FixIf you resolve an incident while the monitor is still failing, we now stop the repeat "still down" alerts instead of sending them indefinitely, and warn you at resolve time that the monitor is not actually back yet.

• ImprovedWhen a site returns 403 (blocked) or 429 (rate-limited) to our checker, the incident now says it is blocking or throttling our checks (and is likely still up for real visitors) instead of reporting the site as down. Blocked checks usually mean a WAF or bot protection needs to allow our IPs.

• FixAverage response time now measures successful checks only. Previously, failed checks (a fast block page or a timeout) could skew the number, making a blocked or down site look fast or artificially slow. Charts, the response-time card, and daily summaries are now consistent.

• ImprovedThe crawler reference page now explains the difference between our uptime checks and our visual (screenshot) checks, and gives a single paste-ready Cloudflare rule that unblocks both. Visual checks run a real browser, so a User-Agent rule alone never let them through; allowing our IPs is the reliable fix.

• FixUptime charts now leave a gap for periods with no checks instead of drawing a misleading 100%. Monitors on longer check intervals no longer appear healthy during stretches that were never actually measured.

• FixA failed subscription payment now sends a single notification email instead of two near-identical ones.

A monitor that is completely down now reports its real uptime everywhere, instead of a misleading healthy-looking percentage.

• FixMonitors with 0% uptime could display a high uptime percentage on the dashboard, the monitor detail page, and status pages. A fully-down monitor now shows its true 0%.
• FixAlert rules with an uptime condition now fire correctly for fully-down monitors. Previously a monitor at 0% uptime was treated as healthy by uptime-based rules, so the rule never triggered.

Every notification (email, Slack, Discord, Teams, Telegram, WhatsApp, PagerDuty, Opsgenie, webhooks, browser push, Prometheus) is now produced from a single canonical event, so the facts can never drift between channels.

• ImprovedAlerts now carry identical facts on every channel. Previously the same incident could read differently in email than in Slack or Discord (different root-cause wording, missing linked-server context, or a misformatted metric).
• FixServer load alerts now report the load average correctly, with core count and true percent of capacity (for example "Load 1.9 / 4 CPUs (48%)"), instead of presenting the raw load average as a percentage.
• ImprovedAlert rule notifications now lead with the affected monitor and state what happened (for example "Response time 870ms exceeded the 500ms threshold") instead of only repeating the rule name. Recovery messages name the metric that returned to normal.
• NewNew "Baseline updated" notification: accepting a new visual baseline now notifies your configured channels, so your team knows future visual checks compare against the new reference.
• ImprovedAlert emails now include a complete plain-text version. Text-only clients, previews, and watch notifications show the full alert instead of a one-line summary.
• ImprovedEvery channel type now receives all alert categories it is configured for. Previously some alert types silently skipped certain channel types.
• FixThe overview page now shows paused monitors in their own count instead of listing them as down.
• FixServer load alerts no longer fire separately while a wider outage alert for the same server is already active.
• FixPagerDuty channels configured with older credential field names now deliver across all alert types.
• ImprovedRecovery notifications show the actual recovery time and reference the exact alert they resolve, keeping alert threads tidy in channels that group related messages.
• FixTesting a WhatsApp channel now reports its real delivery status instead of always showing success.

• FixFixed the Screenshots tab appearing empty for monitors whose most recent capture was more than a day old. The gallery now shows your full set of retained screenshots again, regardless of when they were taken.

• FixFixed an issue where accepting a new baseline for a visual check could revert shortly after, causing the same visual-change alert to reappear. Accepting a baseline now persists reliably and clears the incident for good.
• ImprovedHardened visual baseline handling so a baseline is never replaced by an out-of-date screenshot, including for multi-page (sub-page) checks.

• NewWhen we ship an update while you have the dashboard open, a small prompt now offers a one-click refresh so you always run the latest version.
• ImprovedThe date-range pickers now show dates in your organization timezone, matching how times are displayed elsewhere in the app.

• NewView your Uptime and Performance metrics over any date range with a new unified date picker, including quick presets (24h, 7 days, 30 days, 90 days, a year) and All Time.
• NewA new bulk metrics API endpoint lets you fetch uptime or response-time history for up to 100 monitors in a single request, so dashboards you build no longer need one call per monitor.
• ImprovedUptime and response-time history now support 90-day and 1-year windows, and date filtering extends to the DNS, visual, and content history views.
• ImprovedServer resource charts now clearly indicate their 7-day retention window.

• FixSigned-in visitors are now recognized immediately on public pages: the dashboard link appears right away, and opening the sign-in page takes you straight to your dashboard instead of pausing.
• ImprovedThe visual screenshot comparison view now loads faster and without layout flicker.
• ImprovedEnhanced security hardening.

• FixThe in-app guidance that appears when your site's bot protection blocks our visual (screenshot) checks now shows reliably for every affected monitor, with the exact IPs and steps to allow our checkers.

• NewIf your site's bot protection blocks our visual (screenshot) checks, the Visual tab now shows exactly how to allow our checkers, including our check IP addresses and a setup guide, so you can fix it yourself without contacting support. Your uptime monitoring is never affected by this.
• ImprovedPublished our monitoring check IP addresses on the /bot reference page so you can allowlist them in any firewall or WAF.
• ImprovedBehind-the-scenes improvements to how we measure where new signups come from, so we can keep making the product easier to discover and adopt.

• FixFixed an issue that could stop the website chat assistant from replying.
• ImprovedReduced background noise in our error monitoring so real issues surface faster.

• ImprovedWhen a card payment needs attention, your billing emails and billing page now explain exactly what happened (for example, a card that needs verification, has expired, or was declined) and what to do next, instead of a generic "payment failed" message.
• FixCard verification at signup and trial start is now handled separately from renewal payment issues, so a verification step at checkout no longer looks like a failed subscription charge.

• ImprovedGrouped server incidents now separate shared infrastructure impact from monitor-specific issues such as domain, DNS, SSL, visual, content, and API checks. Incident pages show live cluster impact, peak/current counts, and separate open issues instead of leaving stale grouped incidents open.
• FixServer-cluster recovery now closes grouped incidents once the shared infrastructure signal has recovered, even if a separate domain or visual issue remains open on the same server.

• FixRemoved legacy microdata on visible testimonial cards that was causing rich-result validators to flag false-positive errors. The visible testimonials remain unchanged; canonical review markup continues to live in the pricing page structured data.