Alerts
Tells the right person, once.
One alert per problem, in plain words, to the person on duty – grouped with what belongs with it and quietened once it is fixed.
XovrisOn its own
- A lit square is a feature people use today
- Newest: Prove a page reaches a person: run a drill on any escalation policy
Every feature in Alerts, strongest first.
What fixes and prevents a problem comes first. Type to find one, or open a plan to see only its own.
Nothing here matches those words yet. Xovris notes the words, with no address and no cookie, so what is missing gets written.Ask Xovris instead
Lit: live and workingOpen: Xovris is still working on it
See who triages new incidents now
The incidents board says who new incidents go to now, from your on-call rota, and draws the next 24 hours of cover across every rota with any gap flagged.
Follow an alert's route as it goes out
The alerts page shows who is told, in what order and by what, step by step, live while an alert goes out.
Write an alert rule in a sentence
Say how alerts should be routed or quieted in your own words; Xovris reads it back part by part, shows what it would have changed over the last 30 days, and refuses a rule that would hide a critical alert.
Slack alerts say plainly when they are not available
When Slack alerts are not set up, the card says so in plain words and that your alerts reach you by email meanwhile - never a server message.
Alerts that belong together become one incident
Related alerts are merged into one incident, and each merge says why in words.
Objectives for every customer journey, with burn-rate alerts
Each journey you promise has its objective checked every round, with multi-window burn-rate alerts before the budget runs out.
An incident keeps one thread whatever happens
An incident stays in one chat thread through a token change, a provider outage or duplicate alerts.
Every answer shaped for the channel it leaves by
An answer sent by text, chat or email takes the shape that channel needs - including channels Xovris had no built-in connection for.
Webhook alerts secured with one press
On the delivery page, each webhook's signing secret and HTTPS are put right by Xovris with one press each.
The person who asked hears first
When a job waits on someone, the person who asked is told first and alone for ten minutes, and can press the step straight from their phone.
Alerts by any chat service you use
A chat service Xovris has no built-in connection for is connected from its own API, proved with your own test message, and then used for your alerts.
Pages on your phone, and decisions answered from the notification
A phone subscribes once, and a critical page arrives encrypted to that phone and signed by Xovris. A decision names exactly what an Approve approves, the lock screen keeps private details private, and you can answer from the notification. Xovris promises only what your phone’s system allows: an iPhone is told it needs Xovris on its Home Screen (iOS 16.4 or later) before notifications can arrive, and an older iOS is told to update first.
One test proves your alerts arrive, confirmed by the service that delivered them
A test alert goes down the real path and each channel shows its own service’s acceptance – the email service’s message id, Slack’s timestamp – before the alerts step counts as done. A Slack token revoked in Slack shows Slack’s own refusal.
Told in plain words the moment Xovris loses sight of something
When a check’s results stop arriving, it moves to No data and a message says what cannot be seen, for how long, and that nothing there is reported as healthy in the meantime.
Xovris reconnects Slack itself when a Slack workspace is no longer connected
If Slack drops or forgets the connection, Xovris now offers to connect it again for you: it opens Slack's own approval page asking for exactly the access alerts need, reads the new connection back and sends a test to the channel, with Slack's answer shown. Opening Alerts and connecting it yourself stays beside it.
Warned before something you rely on breaks – and one alert for one problem
When a release of something you run turns out to break, Xovris warns you within minutes if you are on that release. It warns you two weeks or more before a certificate renewal would fail, with the cause, so there is time to put it right. And alerts that describe the same incident now arrive as one, not as a stack.
Your alert rules can call the apps you use, and each app’s answer is on the alert
Link an app’s alert action to a rule and, within seconds of a matching problem opening, Xovris sends that app a signed message; if the app does not answer it tries twice more without ever sending a second copy. The app’s answer and how long it took are written on the incident’s timeline, and Settings > Integrations shows each action’s rules, its last answer and its usual answer time.
Write who hears about what in a sentence, and see what it would have changed
Write a rule for sending or quieting alerts in your own words. Xovris checks it part by part, names anything it did not understand, shows each alert from your last 30 days the rule would have changed, and refuses a rule that would hide a critical alert.
Every alert you set up is test-fired the moment you switch it on, and every month after
Switch on an alert and Xovris sends its first test at once, tests it again every 30 days and checks with the provider ten minutes later that the test really arrived. If a test reaches nobody, or the provider reports it failed, you get an alert about that straight away – so an alert that would stay silent on the day you need it is found long before.
One page per cause, and your weekly letter shows exactly what was held back and why
When a single cause sets off hundreds of alerts – a domain whose DNS goes down, say – you get one incident and one page, with every alert kept under it. Alerts that keep clearing by themselves wait a few minutes before paging, and critical ones still page at once. Your weekly letter lists every such change with the evidence behind it, what was folded under a cause and what cleared without paging anyone, and a suggested change is made within your policy, or kept on Alerts where your policy leaves it to you.
A status update for your own customers, drafted for you and posted only when you say so
When something breaks, Xovris drafts the update for your status page and your subscribers in plain words. Nothing is posted until a person with the right to do so presses Post, and the server keeps out every host, address, path or internal id whatever was typed – the update says how many details were held back and why.
A quiet week ends with an all-clear that proves your alerts still reach you
When you have had no alert for 7 days, Xovris sends one all-clear through your usual route with its latest test result, once a week. If your usual channel is broken it comes by your fallback and says why, so silence always means all is well, never that alerts stopped arriving.
See how many of your alerts were worth waking someone for
Alerts shows your own alert quality for the last 30 days: how many alerts paged someone, how many of those a person acted on, and how long it took a person to respond – the usual time and the slowest one in ten – with the alerts behind every figure. Before anything has paged it says so, instead of showing a zero.
Teams, Discord, PagerDuty and more set up in one step, the test seen arriving
On Alerts every channel sets up from its own row: give it the address or key, press once, and Xovris adds it to the route every alert takes and sends a test. The test shows the provider’s own word – delivered, accepted, or failed with why – and changes only when the provider reports it. Each channel also says what it costs, whether it confirms arrival, whether you can act from the message and whether it reaches you through Focus.
Every alert on your own clock, with the exact moment in UTC for whoever checks the logs
Alert emails, Slack messages and texts give the time on your own clock, right on both sides of a clock change, and the same moment in UTC to the second for anyone comparing it with a log. Every alert email is checked for accessibility, and a screen reader reads it top to bottom, just as the eye does.
Every alert arrives on time and complete, even if Xovris’s AI is down
An alert is sent first and explained after. If the AI is slow or unavailable, the alert still goes out with its seven facts exactly as written, and anything added later must point to something in this incident or it is left out.
Each alert tries the channels that reach you, cheapest first, and stops when you answer
A page goes to your own verified channels in order – free ones first, then the ones that interrupt least, a phone call last and never within the first three minutes. How long each step waits comes from how quickly you usually answer, every step says why, and it all stops the moment anyone acknowledges or you reply stop.
Critical alerts go out in under a fifth of a second, even when Xovris is busiest
Measured with a hundred investigations running at once, 99 in every 100 critical alerts reached the delivery provider within 194 milliseconds – against 182 milliseconds when Xovris was quiet.
A renewed certificate is proven, and you are not woken for it
When a certificate renews, Xovris proves it – a new certificate with a later expiry, seen from every place Xovris checks from – and tells you in your weekly digest instead of paging you. If no renewal arrives in time you are told while there is still time to act, and a certificate issued by a company you never allowed is raised at once.
A critical alert reaches you even before you choose how to be told
If something critical happens before you have set up a channel, Xovris still reaches you – by email to the address you sign in with and on any device with notifications switched on – saying why it came that way and where to choose your channels. Every page also lands in your Xovris inbox.
Every email from Xovris reads complete as plain text, in every major mail app
Every email Xovris sends now carries everything its designed version does in its plain-text part too – every link, every heading and the words on its buttons, signed off with the site’s address – so a screen reader, a text-only mail app or a watch misses nothing. Each email is checked on every build against thirteen families of mail apps, from Gmail, Outlook and Apple Mail to Yahoo, Proton, Fastmail, HEY and Thunderbird, in light and dark.
Pages when a key journey’s target is really at risk, with no rule to write
Give a journey a target and Xovris pages you only when it is failing fast enough to matter – checked over a short and a long window together, so a blip never wakes you – and opens a ticket for a slow drift. It keeps its place through a restart.
Every alert says what happened and the one thing to do next
An alert from Xovris now tells you what went wrong in plain words, what changed just before it, what Xovris is already doing about it and the one action worth taking – the same in email, chat and on your phone. Privacy mode keeps the details out of inboxes and shows them only once you sign in.
Every message from Xovris starts with your own phrase, so a fake one gives itself away
You can now choose a few words only you would know and every email and Slack message from Xovris starts with them – a message without them did not come from Xovris. Each one also says that Xovris never asks for a password, a sign-in code or card details, and its buttons only open a page where nothing changes until you press there.
Alerts say when Xovris has seen and fixed this exact failure before
When something fails the way Xovris has seen before, the alert says so – how many times, since when, and how often the fix worked – never naming anyone it happened to. When it is new, the alert simply says nothing about it.
A runaway burst of messages can never crowd out your alerts
Every workspace has a generous hourly allowance on each channel – 500 emails, 500 Slack messages and 2,000 webhook calls – far above real alerting, because related failures already arrive as one incident. A message over it is never dropped: it waits for the next hour, Alerts says so in plain words, and your other channels keep sending as normal.
Alerts that say what happened, why, and what is already being done
Every alert carries the exact error, what it means, the cause once found and what Xovris is doing about it; a test alert shows each provider’s own acceptance, built-in email is one press, and a failed channel falls back to email and says so.
Alerts that say where things stand, what Xovris has done and what comes next
Every alert now opens with where things stand – Handled, Fixing now, Ready for your review, Needs one tap and the rest – and says Handled only once Xovris has checked the fix worked. Each one carries the exact error, what it means, the cause and how sure Xovris is, who is affected, what Xovris has done, one thing to do and when the next update comes, on email, Slack, push and webhooks. When Xovris acted on its own, the alert names the setting that allowed it, with a link to change it.
Where alerts go, chosen with the facts in front of you
Before you pick email, Slack or a webhook, Xovris shows what each costs you, whether it tells you the alert was really delivered, whether you can act from inside the message, and whether it works in your country – and where a channel does not reach yet, the cheapest one that does.
Every check alert quotes the failure the way you would see it, and where it was seen
When a website, certificate, domain, page, API or journey check fails, the alert now opens with the failure in plain terms – HTTP 503 from your checkout page, the domain not found by the resolver, the certificate with 6 days left, the page that loaded without the words it must show, the step of the journey that failed – and the places it was seen from, never a code only Xovris understands.
Every email from Xovris looks and reads like xovris.ai, in any mail app
Alerts, updates, digests, fix reports, the weekly letter, sign-in codes, status-page notices, account emails and the notice you get if Xovris’s own alert delivery is ever interrupted now share one design: the Xovris mark and name drawn as text, so they show even with images turned off, a preview line in your inbox that says what happened, dark mode in Apple Mail and Outlook, and the same layout in Outlook for Windows. No email carries an image or anything that reports when you open it, and each still has a plain-text version complete on its own.
Every alert shows the exact error your checks saw
Uptime, certificate, DNS, page and API alerts now carry the exact failure the checks recorded – the error code and message, or the status a failed check answered – beside what it means in plain words, the cause or that it is not found yet, what Xovris is already doing and the one thing to do, on email, Slack, webhooks and the app alike.
On-call in Xovris: rotas, escalation, cover in a sentence and leave that never breaks the rota
You can now set up who is on call – daily or weekly turns, office hours, one-off swaps – in your own time zone, so a 9 am handover stays 9 am when the clocks change. If nobody acknowledges, Xovris pages the next person on time. Trainees can shadow real pages with a note on what Xovris did and why. Say “cover me Friday 6–10 pm” and Xovris asks only people who are free, rested and awake, and changes the rota when one says yes. Time off from BambooHR, HiBob or Personio and each person’s public holidays are read in, and any gap in the next 14 days is flagged with who is away. Pages are shared fairly, weighted for night work, with the reason for each.
Stop the weekly letter in one click – and see what Xovris held back
Every weekly letter carries a one-click unsubscribe your email app can show beside the sender, and a link that asks before anything changes, with an undo. The letter also says what Xovris held back that week and why – each count opening the alerts behind it – so a quiet week is shown as quiet, never as nothing checked.
Opening an alert email never counts as handling it
Only pressing Acknowledge on the alert’s own page counts. Alert emails carry no tracking pixel, and the link checks that mail apps and privacy proxies run in the background change nothing – so nobody stops being paged because their phone opened an email by itself.
Tell Xovris once when something is normal, and it stops waking you for it
You can now say it in your own words – “the shop is closed for maintenance on Sundays” – and a check that fails inside that window is recorded, but nobody is paged. If it is still failing when your window ends, you are alerted at once and the alert says why it waited. Ask Xovris to forget it and the next failure pages you as before.
Say what to watch, and see its alert proved before it starts
You can now say what to watch – a certificate, a page’s words, a slow site, a job that must run – and Xovris checks it as you type, shows the exact alert you will get, sends it as a test to every channel you have and switches it on. Each watch keeps its state in place: what it saw last, when it looks next and everything it has seen.
Every number in your weekly letter points at what it counted, and one click stops the letter
The weekly letter says what Xovris checked, found, fixed and held back – and each figure links to the records it counted, so nothing in it is a guess. Your mail app’s own Unsubscribe button stops the letter alone, never your alerts, and you can turn it back on from the page.
Slack alerts, and a disconnect that proves the access is gone
Alerts post to the Slack channel you choose. When you remove Slack, Xovris revokes its access and checks with Slack that the access really is gone before saying so.
Approve, change or decline a fix from the alert itself, on your phone
The alert’s link opens the exact fix Xovris has prepared, with no sign-in: approve it, ask for a change or decline it. Your role is checked at that moment, a forwarded or reused link is refused, and your decision lands on the incident’s timeline straight away.
Prove a page reaches a person: run a drill on any escalation policy
Each escalation policy on On call has Run a drill. It pages that policy’s people with a test incident marked as a test – sent exactly the way a real alert is – and its row shows the drill as it runs: every level, who was paged, by which channel, when it was queued, accepted, delivered and acknowledged. It passes only when a person acknowledges inside the policy’s minutes; when a level is not answered the next level is paged, just as in a real incident, and a channel that never reported delivery is named so it can be fixed. Drills never count in acknowledgement-time figures.
Code alerts go to the person who changed the code
Send an error to whoever wrote the change that caused it: Xovris pages the suspect commit’s author first, then the owners your CODEOWNERS file names for the failing file, and otherwise your usual route – and each page says which rule chose it.
See who is on call, and every gap a fortnight before it opens
On call shows who answers right now and over the next day, drawn from your rota with leave and public holidays applied. Every stretch with nobody on call in the next 14 days is named with its reason and filled in one step. Ask for cover in a sentence – “Cover me Friday 6-10 pm” – and only people who are free, rested and awake then are asked; the first yes changes the rota. Each escalation is drawn as the pages it will send if nobody answers.
Move from Opsgenie in one step
Give Xovris an Opsgenie API key (US or EU) or its export, on the business plans, and it brings across your people, teams, schedules with their overrides, escalations, routing rules, alert sources, heartbeats, maintenance windows and each person’s text and call steps – first as a plan you can read, then for real, every piece read back and undoable. Before you switch, Xovris sets its own rota beside Opsgenie’s real alerts of the last week and names every page that would have gone to someone else; when you say so, it switches Opsgenie’s own alert intake off itself, with one press to switch it back.
Your site is checked for broken layouts every day, and only real faults alert you
Each site you add is looked at every day on phones, tablets and computers, twice a moment apart so a page still loading is never mistaken for a broken one. Text cut off, a button too small to press or something covering your checkout alerts you once, and again when it is fixed. Small polish points are kept in a list for you to read when you like and never wake anyone.
The other areas
Words on this page, in plain English
Each technical word this page uses, said simply.
- A message sent to your own system (webhook, webhooks)
- An automatic message one service sends to an address you choose the moment something happens.
- Secure web connection (HTTPS)
- The encrypted way a browser and a website talk, so nobody on the way can read or change what passes – the padlock in a browser.
- Program-to-program connection (API, APIs)
- A way for one program to ask another for information or to do something, without a person clicking through a website.
- Domain Name System (DNS)
- The internet’s address book: it turns a name you type into the number of the computer that answers for it.
- Web connection (HTTP)
- The way a browser asks a website for a page and the website answers.