Skip to content

What does a check look at, and how do I change it?

Every monitor has an Uptime check. It opens the address every 5 minutes and counts it as up when the page loads with a success status within the time set under Gives up after (30 seconds for a new monitor). With a keyword, the page must also contain the keyword, or not contain it, as you chose.

Read a monitor's page

  1. Open Monitors in the side menu.
  2. Click the monitor.

At the top, an icon beside the name shows the monitor's state (see What the colours mean), and one sentence says for how long, since when and what it means, for example Down for 8 hours, 12 minutes, since 01:14. The server does not answer. When there is something to look at, What our checks saw gives one line per place with its time, then only what that evidence supports: a cause and what you can try, or, when every place got no answer in time, the question Does the site open for you? with what each answer means. Check again now asks our places to check it at once, without waiting for the next check; anyone in the account can press it, once a minute. Until every place has answered, for up to 2 minutes, the button reads Checking now; then the line under the box says when it was checked and what the monitor shows now. Their answers update the lines and can turn the monitor Down or Up, but they do not change the uptime figures. Open the site opens it in a new tab, and Details shows each place's result, status, error code and answer time; quote the error code when you contact us.

Below it, four numbers, and a fifth when the SSL certificate is checked:

  • Last 24 hours and Last 30 days: the share of results that found the website up, and how many outages there were. Hours in a row when the website was down count as one outage. Results that need attention count as up.
  • Answer time: how long the website took to answer, on average, over the last 24 hours.
  • Checked: how often the address is checked.
  • SSL certificate, for an https:// address with the check on: how many days the certificate has left, until when it is valid and who issued it; amber at 5 days or fewer, red with a problem. See How does the SSL certificate check work?

Answer time below them draws the last 24 hours, one point per hour; a red mark under an hour means the website was down in that hour.

On the right, What we check lists each check with a switch. Owners and Admins switch a check on or off in one click; the change reaches our checking servers within a minute. Settings at the top opens the monitor's settings. How the checks are set up further down shows each check's settings: Check interval, Healthy when (for example Status 3xx, 418), Follow redirects when it is off (Off: a redirect is the answer), Gives up after and Keyword.

Hosting

Hosting on the right says who hosts the website. Your own answer is what counts: it shows as a tag, and Owners and Admins change it there with Change. Before you answer, it shows what we can see:

  • Behind Cloudflare, host unknown: Cloudflare (or another service in front, such as Sucuri) hides the server behind it, so nobody outside can see who hosts it.
  • Probably Hetzner, from its address: the website's address belongs to that company's network.
  • Probably Kinsta, from the website's headers: the website's replies to our checks name its host. With a service in front, it says so: Probably Kinsta, from the website's headers, behind Cloudflare.
  • Probably GoDaddy, from its nameservers: the domain's nameservers belong to that company. This is often the company the domain is registered with, which may not be the one that hosts the website.
  • Host unknown: we could not see it.

Below that, also after you answer, two lines about the domain when we have them: Nameservers, the servers that answer for the domain, and Registrar, the company the domain is registered with and the date the registration expires. Both come from public records: the nameservers from the domain's DNS, read once a day; the registration from RDAP, the newer form of whois, read about once a week. Domains ending in .io, .co, .me, .us, .eu, .de or .ch, and some others, have no public RDAP service we can read, so they show no Registrar line.

We ask once: Is that right? for a guess, Where is it hosted? otherwise. Type to search the list of hosts, which also holds the hosts you use on your other monitors; for a host not in it, type its name, choose Add "its name", then Save. Skip stops the question for that monitor; Add or Change is still there, and Not set in the list removes the host. Choose the company that runs the server, not a service in front of it: every website behind that service would share one alert email.

Monitors with the same Hosting and no Server that go down together come in one email (one email for several monitors of one host).

Read the last result

Every address is checked from more than one place, one place at a time: the places take turns, so your website gets one visit per check interval. Each check moves to the next place about every half hour, or every two of its intervals when it is checked less often than every 15 minutes. Under Last result, folded until you open it, a line says that the places take turns and, for each check, where it is checked from now; you see one line per place. On port 22 (SSH) the line says Checked from one place only: our other places do not check this port. (Down only when a second place agrees):

  • the place it was checked from, for example Amsterdam;
  • the result, in words;
  • the status the website answered with, for example Status 200;
  • how long ago it was checked. Point at the time to see the exact date and time in your own time zone;
  • where the check ended, when another host answered: Redirected to example.org: we checked that page., or Ended on auth.example.com, a sign-in page.

A place whose turn has ended shows its last result in grey, with its time and the words "Its turn has ended": once the next place's result is more than one check interval later, it no longer sets the monitor's state. A result that has not been renewed for a while is grey too, with "No recent result".

The results refresh by themselves every minute while you have the page in front of you and are using it.

Result Counts as What it means What to do
Up Up The page loaded as expected Nothing
Slow Needs attention The page loaded, but slowly Look at the server's load or the page's size
Server error (500), Page not found (404) and other error statuses Down, once a second place agrees The page answered with another status than expected; the number is the status Open the page and look at what it shows; What our checks saw on the monitor's page says what probably happened
Asks for a password (401), Access refused (403), Too many requests (429), Moved elsewhere (301) and other redirects Needs attention The server answers, but not with the page; visitors may well get the website What our checks saw on the monitor's page says what probably happened and what you can try
Unreachable Down, once a second place agrees The website did not answer at all Check that the server is running and the address is right
SSL certificate problem Down, once a second place agrees The certificate is expired, not trusted or for another name; visitors see a warning Renew or fix the certificate; the SSL certificate check says how for each problem
Keyword missing or Unwanted keyword found Down, once a second place agrees The keyword rule did not hold Check the page, or change the keyword
Keyword not checked Needs attention The page is larger than the 1 MB we read, or came compressed in a way we could not read, and the keyword was not in the part we read; no alert Choose words near the top of the page, or watch a smaller page (keyword)
Behind a sign-in page Needs attention The check ended on a sign-in page of the website's own domain, so the keyword could not be found; no alert Watch a page anyone can open without signing in, or remove the keyword (Behind a sign-in page)
White screen (500, empty page) Down, once a second place agrees The server answered with an error and nothing on the page, so visitors see a white screen Undo the last update, check the error log, or ask your host; What our checks saw on the monitor's page says what probably happened
Empty page, Redirect loop, Too many redirects, the WordPress, PHP and Cloudflare errors Down, once a second place agrees The page loads wrongly or not at all for visitors What our checks saw on the monitor's page says what probably happened and what you can try
WordPress maintenance mode Updating (needs attention) for its first 10 minutes, then down, once a second place agrees WordPress is installing an update; after 10 minutes it is most likely stuck Wait; if it stays, delete the .maintenance file in the folder that holds wp-config.php (WordPress maintenance mode)
Private or internal address Needs attention The address points to a private network, which we never check Use the website's public address
Basic Auth password needed Needs attention The website asks for a Basic Auth user name and password, and none is saved Save them under Basic Auth in the monitor's Settings
Basic Auth password refused Needs attention The website asked for its password and refused the one saved here; it was probably changed Save the current one under Basic Auth in the monitor's Settings
Basic Auth is off Needs attention The Basic Auth warning found the website open to everyone, twice in a row, where it should ask for a password Turn the website's Basic Auth back on
Our check was blocked by a firewall Needs attention A firewall in front of the website stopped our check Allow our checks in your firewall
Could not check Not counted Our checking server had a problem; it is never counted as down Nothing; it usually clears within minutes
A result in grey Not counted No newer result came in for a while; it is the last one known Look again later; the Monitors list says No recent result

Down only when a second place agrees

One failed check never makes a monitor Down on its own. When the place whose turn it is cannot get the website, a second place checks it at once, usually within seconds:

  • Both fail: the monitor is Down, since the first failure. Its page says where it was confirmed from, for example Confirmed from Amsterdam and Newark.
  • The second place gets the website: the monitor shows Checking again, and the failure is not counted as down. For the next 30 minutes no second place is asked about this monitor, so your website still gets one visit per check interval, and its failures in that time are not counted either. If this keeps happening, the website may block visitors from some countries: allow our checks.
  • The second place cannot answer (it is busy, or something on our side or a firewall stops it): the monitor shows Checking again, and it turns Down after 3 failures in a row from the first place. That place is not asked again about this monitor for 30 minutes.

Checking again is not the button Check again now: it is the state while a second place is asked, and it comes by itself.

While the second place is asked, the monitor shows Checking again with a line such as Amsterdam could not reach it at 10:14. Asking Newark.; in the Monitors list the same in short, Amsterdam could not reach it; asking Newark. When no second place can check it, the line says so: It turns Down after 3 failures in a row from Amsterdam.

A port monitor on port 22 (SSH) is the exception: only Newark checks it, so it turns Down after 3 failures in a row from Newark, without a second place. Its Last result says Checked from one place only: our other places do not check this port.

A monitor that is down stays Down until two results in a row from the same place, or from two places, find the website up again: the first good result asks the second place at once, so it usually turns Up within seconds of the website's return. Its page says Up again from Amsterdam at 10:31; confirming. in between. Only failures inside a confirmed down time count as down in the bars and the uptime; an hour that holds only failures not confirmed is a grey bar, Not counted as down.

You get one email when the monitor turns Down, and one when it is Up again (which emails you get, and when).

What the colours mean

Each state has its own colour and its own shape, so you can tell them apart without the colour too. Point at an icon to read its word.

Icon State What it means
A green circle with a tick Up Visitors get the website
A red square with a cross Down Visitors do not get the website: two places agreed, or one place 3 times in a row when no second place can check
A grey dashed circle with a turning arrow Checking again One place could not get the website and a second place checks it; not counted as down. See Down only when a second place agrees
An amber triangle Needs attention Something to fix, but visitors may well get the website; counted as up
An amber triangle Updating WordPress is installing an update; counted as down only if it lasts more than 10 minutes, see WordPress maintenance mode
A grey dashed circle with a question mark Could not check Our side could not tell; never counted as down
A grey dashed clock No recent result No result for longer than expected
A blue clock Not checked yet Just added; the first result is coming
A grey circle with two bars Paused Someone stopped the checks
A grey slashed circle Checks off Every check of the monitor is switched off

Red is kept for down: a red icon means two places could not get the website, or one place 3 times in a row when no second place can check, so visitors probably cannot either.

Read the Monitors list

Each monitor has one row. By default the monitors that are down are at the top, then the ones checking again, then the ones that need attention. Under the page's title, Recommendations holds our tips for your monitors, when there are any (the notices). Above the tabs, a line says how many monitors your plan allows and how many you use, such as "You use 7 of 25 monitors on the Plus plan."

Above the list, tabs show the monitors by their state, each with how many of your team's monitors are in it, whatever the search and the filters show: All, Down, Needs attention, Up and Paused. Click a tab to see only those monitors; the search and the other filters stay as they are, Not paused included. Paused shows the paused monitors only; the other tabs leave that choice out again. Monitors in none of these states are in Filters (on a wider screen More filters), under Other states: Checking again, Could not check, No recent result, Not checked yet and Checks off, each with its count; click one to see those monitors.

The search finds a monitor by its name, its address, its group or its tags. The Group and Tag filters (on a phone in the Filters panel) show only the monitors of one group, or with one tag.

At the end of the tabs' line, for example 99.98% uptime, last 24 hours, is the uptime of all the monitors except the paused ones, over the same 24 hours as the bars. Results that need attention count as up, and it is never rounded up. Point at it, or click the (i) beside it, to see, for example, Down at some point in 4 hours, all monitors added up: for each monitor except the paused ones, the hours with at least one down result, added up. Two monitors down in the same hour make 2 hours, so the number can be more than 24. No uptime data yet means there is no result from the last 24 hours to count.

The list has a column for each of these, named at the top:

  • Monitor: the icon before the name shows the state, as in What the colours mean; the monitor's own page shows the same icon. After the name come its tags as small labels, three at most and then a count such as +2 for the rest. Under the name one line says the state and for how long, for example Up for 4 days, 13 hours, Down for 8 hours, 12 minutes, Basic Auth is off since 09:14. Anyone can open it. or Slow: answers in 3.2 seconds. After it comes the group the monitor is in, if it is in one, with a folder icon; click it to open the group. On a wide screen the group has a column of its own instead.
  • Group: on a wide screen, the group the monitor is in, if it is in one; click it to open the group. On a narrower screen there is no Group column and the group is under the name. When the list shows only one group's monitors, or only monitors in no group, or none of the monitors shown is in a group, the group is not shown.
  • Last 24 hours: one bar per hour, the newest on the right. A green bar is an hour when the website was up, a red bar an hour when it was down at least once, an amber bar an hour with results that need attention (striped), and a flat line at the foot an hour that was not checked or has no results; a day without any results is one flat line. Red and amber bars are taller than green ones, so they stand out by their height too. Point at a bar to see its date, its hours in your time zone, and what happened, for example Down about 15 minutes, 75% up.
  • Uptime (30 days): the share of the last 30 days' results that found the website up, the same number as Last 30 days on the monitor's page. Results that need attention count as up. It is never rounded up, so a single outage is enough to keep it from 100%. No data means there is no result to count yet.
  • Interval: the time between two checks, for example 5 min.

Click a row to open the monitor; its Settings are one click from there.

On a phone each monitor is a card: its box and its kind on the first line with its state on the right, then the name, the group, the bars, and the uptime and the interval at the bottom.

Sort the list

Click the name of the Monitor, Uptime (30 days) or Interval column to sort the list by it: Monitor from A to Z, Uptime (30 days) lowest first, Interval shortest first. Click the same column again to reverse the order; the small arrow beside the column's name shows which way it goes. Monitors without an uptime or an interval come last either way.

Sort, beside the filters, has the orders no column has: Problems first, the order the list starts in, and Slowest first, by how long each website took to answer on average over the last 24 hours, shown under each name while that order is chosen; monitors without an answer come last. Where the column names are not shown (on a phone, in the Filters panel), Sort has every order both ways: Problems first, A to Z, Z to A, Uptime, lowest first, Uptime, highest first, Interval, shortest first, Interval, longest first and Slowest first, and the list says which order it is in. The order stays in the page's address, so a bookmark or a shared link keeps it.

Act on several monitors at once

Owners and Admins see a box before every monitor. Tick the boxes of the monitors you want, or tick the box at the top of the list for every monitor the search and the filters show (a dash in it means some are ticked). On a phone, tick one, then Select all in the bar. To tick a range, tick one box, then hold Shift and tick another. Once one is ticked, a bar takes the place of the search and the filters and stays at the top of the window; it says how many are selected:

  • Move to… and Move put them in a group, or in none: see How do groups work?.
  • Add tag: type a tag (or pick one of yours) and click Add tag, or press Enter. A monitor that has it already, or holds 10 tags, is left as it is.
  • Pause and Resume stop and restart their checks, as on each monitor's page.
  • Group alerts, then Switch on or Switch off, for their group's addresses (Group alerts for many monitors).
  • Interval… and Set interval give every check of the selected monitors one check interval. Only the intervals your plan includes are offered; a monitor already at it is left as it is, and a paused one keeps it for when you resume it. See Check interval. Without an interval chosen it says Choose an interval first. It changes nothing when a monitor's check gives up after longer than the interval chosen (Gives up after is longer than 30 seconds on 2 monitors. Not changed.): choose a longer interval, or shorten Gives up after in those monitors' settings. Undo asks the same again, and what your plan includes, before it puts the intervals back; when it cannot, it says why and changes nothing.
  • Gives up after… and Set timeout give every web check of the selected monitors one time to wait for an answer: 10, 20, 30 or 45 seconds. A port or ping monitor has none and is left as it is, and so is a monitor already at it. Without a time chosen it says Choose a time first. It changes nothing when a check would wait longer than its own interval (45 seconds is longer than the check interval of 1 monitor, 30 seconds. Not changed.): choose a shorter time, or a longer interval first. See Gives up after.
  • More removes one of their tags, or deletes the monitors.
  • Clear, or the Escape key, unticks them all and brings the search and the filters back; Undo stays.

After each of these the bar says what was done, with Undo to put it back as it was, until you act again or select other monitors. Deleting asks first and names the monitors; their checks, results and history go with them, and it cannot be undone.

Change a monitor's settings

Owners and Admins change the name, the checks, the keyword, Basic Auth, the check interval and the rest on the monitor's Settings page: see How do I change a monitor's settings?.

Pause and resume

Pause a monitor while its website is moved or rebuilt, so it is not checked in the meantime.

  1. Open the monitor.
  2. Click Pause checks under Quick actions on the right. To give a reason, pause it under Stop watching on its Settings page instead; the reason is shown on the monitor's page, so the rest of the team knows why.
  3. The monitor is marked Paused.

To start again, open the monitor and click Resume checks under Quick actions.

Delete a monitor

Under Stop watching on the monitor's Settings page: see Stop watching. Or tick monitors on the Monitors list and choose Delete monitors under More. A deleted monitor moves to Deleted: its checks stop at once and it sends no alerts, it no longer counts against your plan, and its address can be added again.

Deleted monitors

While your team has a deleted monitor, the Monitors list has a Deleted tab after Paused. It lists each one with the day it was deleted, who deleted it, and the day it is deleted for good: 7 days after.

  • Restore brings the monitor back as it was: its settings, group, tags, alert addresses, results and history. Its checks start again, unless it was paused. When your plan has no room for one more monitor, or the same address is watched again by another monitor, Restore says so and changes nothing.
  • Delete forever asks first, then deletes the monitor with its checks and results at once. Your team's History keeps its lines, with one more saying it was deleted forever, and your alert log keeps the alerts that were sent for it.

After 7 days in Deleted, a monitor is deleted for good. Deleting your team's account deletes its deleted monitors for good at once. Owners and Admins can restore and delete; Members and Viewers see the tab only.