• Live Activities covering several monitors on one server now end when the whole server recovers. • Delete a monitor in Uptime Kuma while it's down, and its Live Activity now ends as soon as KumaAlert next connects. • Live Activities end on time more often: the app retries when it can't reach our servers.

Sync is now simpler. Sign in on any iPhone, iPad or Android device and your servers and settings are there, with no codes to compare and no approvals. Synced data stays encrypted, unlocked by your Apple or Google sign-in. Prefer only your devices to hold the key? End-to-end encryption is now an option under Settings > Account > Sync encryption, with a recovery key so you can always get back in. Existing accounts see a one-time notice and can keep end-to-end. Update KumaAlert on all your devices: once any device moves the account over, older versions can no longer sync changes. Also fixed: - Sync can no longer get stuck waiting for an approval nobody can give. - Sign in with Apple now says why it was refused, instead of always blaming another Apple ID. - "Sign out and remove sync key" is gone. Plain sign out is all you need now.

• A full security review hardened sync pairing, webhooks and stored secrets. Update every device you sync: pairing now uses a longer code. • Crash reports are sent only in the Full privacy mode. • Live Activities end when a monitor recovers, even with recovery alerts off or snoozed. • Fixed a rare freeze at launch, and the app does less work on the main thread in the background.

Fixed: • The Incidents tab no longer freezes briefly on devices with a long outage history. • The Monitors screen could freeze until iOS closed the app while several servers were reconnecting at once. It no longer redraws on every reconnect attempt. • The iPad sidebar no longer redraws every time a monitor changes status, which could freeze the app for a moment after it came back from the background. • An acknowledged alert could fire again later in the same outage, for example after its Live Activity was dismissed, or as a delayed alert that had been queued before the acknowledgement. An acknowledgement now holds until the monitor actually recovers. • Certificate expiry alerts could arrive many times a day when a site is served with more than one certificate. The same alert now arrives at most once a day. • On some iPads running iOS 17 the app could start without its saved servers and show a store error; it now always opens your data. • Closing the Manage screen could freeze the app until iOS closed it. • The Overview screen no longer re-reads your server list every three seconds to refresh the average response time.

New • Full-screen support for iPhone Duo, folded and unfolded. Fixed • The Overview screen redrew itself twice a second while connected, which could freeze the app until iOS closed it. • The outage timers no longer do their refresh bookkeeping on the main thread, which removes another source of the app freezing. • The main screens no longer read the old pre-4.0 monitor database every time they redraw, which could freeze the app on launch until iOS closed it. • The average response time on the Overview no longer restarts its sampling on every status update.

Fixes in this update: - Maintenance windows show at the right time: one-off windows set in Uptime Kuma, or on a device in another time zone, now use the window's own time zone, and editing one no longer moves it. - Recurring maintenance windows (for example daily 02:00 to 05:00) show their schedule and next window, and saving keeps the schedule. - Setting up the KumaAlert webhook straight after connecting no longer adds a second copy that sent every alert twice. - Fixed freezes when a server with many monitors sent an update, and on the iPad Overview response-time chart. - Fixed the app closing at launch on some iPads running iOS 17, when a snooze ran out on the Incidents tab, or when deleting a custom header. - Tapping a notification or widget now opens the right server's monitor, and iPad pause and edit act on the right server. - The home screen widget asks you to open KumaAlert instead of showing sample monitors before the app has synced.

Bug fixes - Critical alerts could arrive as Time Sensitive after push notifications were re-registered, even with Critical Alerts switched on. - Acknowledging or resolving an incident on a device that had been unlinked from its account failed with a bare "HTTP 409". It now says the device needs relinking and where to do it. - Re-registering push notifications while signed in unlinked the device from its account, which paused sync and incident actions without saying so. The device now stays linked. - A device that loses its account link now shows a sign-in prompt on the Overview screen, not only when it has no servers.

Added: • Privacy modes in Settings › Privacy: Full, "Local only, keep push", or Local only, where nothing is stored on KumaAlert's servers. Setup asks which you want. • Delete your account from the app, in Settings › Privacy. Changed: • KumaAlert no longer sends crash reports.

New - Overview now warns when Uptime Kuma is sending a server's alerts to another device, such as your previous phone, and one tap points them back at this one. - When you're signed in, KumaAlert moves each server's webhook to your account's address, so alerts keep arriving after a phone swap or reinstall. Fixes - Notification settings no longer show push as Active for a server whose webhook still posts to another device. - Rotating a server's webhook secret updates Uptime Kuma again, instead of rotating the secret and leaving alerts paused.

• Long-press a maintenance schedule in Manage to pause, resume or delete it without opening it. • With more than one server, the Manage tab's All view no longer shows one maintenance schedule twice while hiding another, and editing, pausing or deleting a schedule there now acts on the right server. • iOS connects to Uptime Kuma servers behind proxies or tunnels that rewrite the Host header, which Uptime Kuma's WebSocket origin check used to reject with HTTP 400.

Improved account sync and notification reliability. Fixed widget status, paused monitors, and background refresh.

- A connection blocked by Cloudflare Access or a proxy now says why. When "Test Service Token" passed but Connect still failed, the app only showed a generic "Could not connect" message. Connect now shows what actually refused the connection, and the test uses the same connection method as Connect, so a pass means Connect will work. - Lighter on battery and data while the app is open. On a busy Uptime Kuma server the app synced your monitor status and incident history every few seconds. It now syncs when a monitor changes state, you connect, or you open the app, and otherwise at most once a minute.

Fixed: • Linked Devices shows each device's current version. A device could stay listed on an old version after updating. It now refreshes its details once a day. • A quick recovery cancels a delayed alert. On monitors set to alert only when down, an outage that ended inside its alert delay could still send a late "is down" notification that acknowledging could not stop. • Alert titles read correctly for plural monitor names.

Fixed: • Incident history stays with the correct account. Switching accounts can no longer let an older sync republish or write the previous account's history, and a sign-out request that fails keeps you linked until it can be retried, instead of half-unlinking you. • Concurrent updates no longer overwrite newer evidence. A delayed snapshot, a start correction or a stale check now defers to a newer observation, so a quick acknowledgement, note or resolution stays visible on every device instead of being undone by a request that was already in flight. • Sync preserves your complete history. Outages sharing a timestamp at a page boundary are no longer skipped, a first sync now includes older history, and a save that fails no longer advances past the rows it never wrote. A stale snapshot can no longer delete newer outages. • Your old on-device history imports safely. The imported rows and the marker saying the import finished are now written together, so a failure can simply retry rather than leaving a partial history marked complete. Outages you closed manually, or that ended because a monitor was paused or deleted, keep that reason instead of being relabelled a recovery. • A pending or maintenance reading no longer splits an ongoing outage in two or invents a recovery. A confirmed outage start survives them. • Monitors that share an ID across two servers no longer replace each other's notifications. • An invalid numeric reading from a server can no longer crash the app.

Fixed: • Incident sync follows live monitor updates. Status and history updates now trigger a relay snapshot followed by a refresh, including a follow-up when another request is already running. This lets missed webhook recoveries reach incident history and replaces pending “Detecting” cards after confirmation. • Long outages use Kuma's transition time. The dashboard's last 20 heartbeats no longer supply an invented outage start. Confirmed transition times can also correct an existing open incident's start on the relay while preserving its acknowledgement and notes. • The app stops locking up while it syncs your incident history. A pull from our servers re-read every incident you have, once for every page it fetched, on the same thread that draws the app — and then swept the whole table again afterwards, every single time. On a large history that was enough to freeze the Incidents and Monitors screens for seconds at a stretch. The re-reads now happen once per sync. • Outage history no longer depends on turning notifications on. Your devices identify themselves to our servers to sync incidents, and that identity used to be issued only as part of setting up notifications, so a phone where you declined the prompt was never recording any history at all. Every device now sets itself up on launch, and wanting a record of your outages is no longer tied to wanting to be alerted about them. • A phone with notifications turned off stops pretending it has your history. Outage history is recorded on our servers, and the app only starts talking to them when you set notifications up, so a phone where you declined the prompt was never recording anything. It showed “All quiet, everything is running smoothly” anyway, under outage cards stuck on “Detecting” that could never resolve and a tab badge that could never clear. That phone now says plainly that it isn't recording, and offers to set itself up. Your other devices were never affected. • The Incidents badge includes monitors awaiting confirmation. Pending cards use the current monitor list, and both phone and iPad badges count them alongside confirmed outages without duplicates.

Fixed: • The Incidents badge stops disagreeing with the screen. The red number on the tab counted every unresolved outage the relay still had open, while ACTIVE NOW hid the ones your server had already reported back up — so the badge could read "2" over a list showing one. Both now count exactly the same thing, as do the per-server chips. • A recovery updates the Incidents screen while you're looking at it. Outages only closed on the screen when the app was reopened or a push arrived, so a monitor that came back kept its group saying "1 still down" and its detail screen saying "Ongoing". A recovery seen on the live connection now refreshes the incident list straight away. • An outage timer no longer restarts when you open the app. On a cold start the app could tell the server "this monitor is down as of right now" before your Uptime Kuma history had loaded, which restamped an outage that had been running for hours. It now says nothing at all until it knows the real start time.

Fixed: • A server can no longer vanish when you add another one. If your account held a server this device couldn't decrypt — which happens after linking, resetting and re-linking a second device — the next sync read that as "deleted elsewhere" and removed it from this phone. A sync that can't read everything on the account is now never allowed to delete anything. • The monitor type badge stops cutting words in half. Docker read "DOCKE", PostgreSQL "POSTG" and SQL Server "SQL S". Every type now has a badge that actually reads as that type. • Scrolling past the response time graph scrolls the page. The graph claimed any swipe that started on it, so on a monitor screen you often ended up scrubbing the timeline instead of scrolling. Press and hold on the graph to scrub; a normal swipe now just scrolls. • The hostname field rejects input that could never be a host. Typing something like 192.168.7.2= created a monitor that then reported a confident DOWN about nothing. Malformed addresses are now refused with an explanation, and the field says "hostname or IP" so it's clear both are accepted. (A name that is well formed but doesn't exist still can't be caught here — only your Uptime Kuma server knows that.) • The app no longer freezes while a Live Activity reports its token. Reporting Lock Screen activity tokens ran on the same thread that draws the app, and with several outages at once iOS could kill the app for being unresponsive. • The Monitors tab no longer stutters during a large outage. It re-read every open outage from the database on every redraw; with a lot of monitors down at once that was enough to lock up the screen. • Choosing monitors for a maintenance window opens instantly on large fleets instead of laying out every monitor at once. • A request to our servers that opens but never answers now gives up after 15 seconds instead of holding on for a minute.

> Items tagged iOS, iPad or Android apply to that platform only; untagged items apply to both. Changed: • iOS: Incidents are now tracked on our servers, not reconstructed on your phone. Your outage history is the same on every device, survives a reinstall, and no longer depends on the app being open when something breaks. Existing incidents on your phone are uploaded once, automatically. • iOS: An outage whose server has gone quiet is marked Stale and shows "no data since 14:32" instead of a downtime counter that keeps ticking on a monitor nobody can actually see. • iOS: Acknowledge, notes and Mark as resolved apply everywhere, not just on the phone you tapped them on. Fixed: • iOS: A group monitor going down now starts its Live Activity and keeps its incident. Group outages were being closed by the app moments after they began, so you got the notification but no Lock Screen activity and no entry in history. • iOS: Duplicated, split and phantom "still down" incidents are gone. Outages that recovered while the app was closed now close properly, and a monitor you delete or pause no longer leaves an outage open forever. • iOS: Generating a diagnostic report no longer freezes the app. Settings > Diagnostics could lock up for several seconds while it gathered logs. • iOS: Per-monitor alert rules stop resetting. Mode, severity, repeat-until-ack and alert delay were being wiped by the app's own background refresh, so a rule you set vanished the next time a notification arrived. • Alert rules now apply on every device signed into your account, not just the one you set them on. Fixed on the server, no update needed. • iOS: HTTPS monitors show their TLS certificate expiry again. The certificate date Uptime Kuma sends was never being read, so the row never appeared on the monitor's detail screen. • Monitors inside a paused group no longer show as Paused. If you organise monitors under group monitors and leave the groups inactive, Uptime Kuma still runs and alerts on the monitors inside them — but reported them to the app as inactive, so every one showed as Paused and Resume did nothing. They now show their real status, and a monitor held inactive only by its group gets a "Group paused" note on its detail screen. • Android: HTTPS monitors show their TLS certificate expiry. Android had the same wire-parsing bug iOS did — the expiry Uptime Kuma sends was never being read, so the row never appeared.

Added • Delete a monitor straight from its edit screen in the Manage tab. • Choose a background for the Overview banner: Gradient, Mesh, Solid or Ink. While everything is healthy the banner takes on your accent colour. • Four more accent colours: Violet, Teal, Orange and Graphite. Fixed • Recovery notifications no longer report absurd downtimes like "1628 hours" for a five-minute outage. • A server you delete stays deleted on every device instead of coming back after the other phone syncs. • Servers keep their identity and credentials when syncing between Android and iOS, and API keys now sync too. • Linking devices merges both server lists without losing either one, and approving a device to sync now works. • Linked devices show the right operating system. • Signing out no longer breaks syncing for your other devices. • Monitors opens on All servers and remembers the scope you pick. • Outages on monitor groups clear themselves again, and you can press and hold anything under Active Now to mark it resolved yourself. • The widget's monitor count matches the app, and it no longer shows zero overnight. • The Dynamic Island stops counting when the outage ends. • Android: Overview totals every server, Monitors starts on All, and outages clear once the monitor reports in again.

This release is mostly speed and reliability. • The lag is fixed. Overview, Monitors and Manage no longer redraw constantly under steady heartbeat traffic, and reordering monitors is smooth again. • A lost connection is now visible. Overview shows an offline banner with a Retry button instead of claiming "live" over stale data. • Fixes for database monitors, maintenance windows, two-factor servers and very large fleets. • Incidents now opens on All servers rather than one. • New: send a diagnostic report to support in one tap from Settings, Diagnostics, and get a short code to pass on. • Signing in with Apple and with Google now lands on one account.

Added: • Snooze alerts for up to a week. The snooze menu on an active incident now offers 30 minutes, 1 hour, 4 hours, 1 day, 3 days and 1 week, so a monitor that keeps flapping can be silenced for as long as the problem lasts. Monitoring keeps running the whole time, and the Snoozed section shows when the pause ends. Fixed: • Per-monitor alert rules stay saved. Setting a monitor to Critical or Down only could snap back to the defaults (All, Warning) the next time account sync ran.

Bug's squashed! Full changelog: https://kumaalert.app/changelog/

Bug's squashed! Full changelog: https://kumaalert.app/changelog/

What's New in Version 4.0 This is a big update. KumaAlert now reaches your wrist, your Lock Screen, and beyond. • Apple Watch app. Your monitors, live outages, and Acknowledge or Snooze right on your wrist, even with your iPhone off. • Severity levels. Mark monitors Info, Warning, or Critical. Critical alerts can break through silent mode and Do Not Disturb, and repeat until you acknowledge them. • Act from anywhere. Acknowledge or snooze from the notification, a widget, the Live Activity, or the new Incidents tab. • A new Incidents tab. Active outages pinned with live counters, downtime stats, and mean time to recovery. • A full family of widgets. Home Screen and Lock Screen, showing only what's down when it matters. • One-tap push setup. KumaAlert creates the Uptime Kuma webhook for you. • Arrange monitors by dragging, plus Siri and Shortcuts, a Control Center control, and per-monitor alert delays. There are also reliability fixes for Live Activities and reconnects, and tougher push security. KumaAlert is now free for registered charities. Full changelog: https://kumaalert.app/changelog/

NEW FEATURES - See how long a monitor has been down. Down monitors now show a live outage duration on the Monitors list, on the monitor's detail screen ("Down for 23 min"), and next to the affected monitor on the Home Screen widget. - Optional "last updated" caption on the widget. Turn it on under Settings → Widget to see how fresh the widget's data is. Off by default. PERFORMANCE - Much smoother scrolling on the Monitors tab. Heartbeat history now lives in memory instead of being written to the on-device database on every check, removing the constant background churn that caused stutters while scrolling. - Snappier everywhere on large servers. The monitor list no longer re-parses every monitor's configuration on each screen update, and incoming status updates are applied on a steady cadence instead of piling up and landing as one big freeze. - The Overview tab no longer rebuilds itself every few seconds. The live average-response figure now updates in place. - The app takes up less space. A one-time cleanup on first launch after updating removes the heartbeat history previously accumulated on your device; history is fetched fresh from your server instead. BUG FIXES - Background widget refresh now works on standard Uptime Kuma servers. The app was calling an endpoint that only exists on some forks and silently giving up; it now uses the standard metrics endpoint your API key is made for, so the widget updates in the background as intended. - History chart tooltips now point at the bar you tapped. Near the edges of the 24-hour chart on a monitor's detail screen, the popup could appear anchored to a neighbouring bar. NOTES Response-time charts and recent checks now stream from your Uptime Kuma server when the app connects, so they may take a second or two to appear after a cold launch and aren't available offline - matching how the Uptime Kuma web dashboard behaves.