Availability
The Uptime Monitoring Availability tab shows your uptime, every outage, what the outages cost, and hours when Google got errors although your site looked up.
The Availability tab of Uptime Monitoring shows if your site answered, and every outage in the date range. It also shows what the outages cost in visits and revenue, and what search and AI bots got while the site was down.
Before you start
- To open the tab, select Uptime Monitoring in the domain sidebar, then select Availability. It is the first tab.
- Lost visits, orders and revenue need Web Analytics. Orders and revenue also need conversion tracking.
- Bot errors and crawl data need Server Log Monitoring. Google's own error reports need Google Search Console.
The uptime card
The card at the top of the Availability tab of Uptime Monitoring has two tabs, and more tabs for the cost of the outages.
Uptime
The Uptime tab shows the share of checks that got an answer from your site, and the total time down. For example, 99.95% with 12 min down, or No downtime. The chart shows the uptime for each hour, day or week. The date range sets the default step.
| Uptime | About this much downtime in a month |
|---|---|
| 99.9% | 43 minutes |
| 99% | 7 hours |
If you compare periods, the tab shows the change against the earlier period. The earlier period shows as a dashed line.
Outages
The Outages tab shows the number of outages in the date range. The note says how long the site took to come back on average, for example Back up after 4 min on average. It can also say Still down or None in this window. The chart counts the failed checks in each step (Failed checks).
Days when bots crawled far less
If Server Log Monitoring is connected, the chart shades the days when search and AI bots crawled far less. Far less means half of the usual crawl or less. The legend calls these days Bots crawled far less. An outage on the same day does not prove that it caused the drop.
What the outages cost
If Web Analytics is installed and the date range has outages, the uptime card of Uptime Monitoring has three more tabs:
- Visits lost
- Orders lost
- Revenue lost
Each figure is an estimate. VitalSentinel compares each outage with a typical same weekday. It takes the visits (or orders, or revenue) of those hours on a typical day, minus what arrived during the outage.
If some outages are not priced yet, the label adds (at least). If Web Analytics has not priced any outage yet, the tab says Web Analytics has not priced these outages yet, so their cost shows a dash.
Orders and revenue need conversion tracking
Orders lost and Revenue lost need conversion tracking, because Web Analytics must see your orders. Without it, the tab says Orders lost needs conversion tracking (or Revenue lost needs conversion tracking). Owners and admins click Set up conversion tracking to open the setup panel for their shop platform. See Ecommerce.
Every outage
The Every outage card of Uptime Monitoring lists the outages in the date range, ten on each page. Click a row to open the outage. These columns show:
| Column | What it shows |
|---|---|
| Kind | A chip with the status code if the site sent one, or else the kind of failure. |
| Started | When the first failed check happened, in your account time zone. |
| Duration | From the first failed check to the next successful one. Ongoing means that the site is not back up yet. |
| Visits lost | Web Analytics: the estimated visits lost. |
| Orders, revenue | Web Analytics: a dash, because Web Analytics has not recorded an order yet. Set up conversion tracking to measure orders and revenue. |
| Bot errors | Server Log Monitoring: requests from verified search and AI bots that got a 5xx error during the outage. |
| Google back | Server Log Monitoring: the full days after the outage until Google and AI bots crawled at 80% of a usual day. |
The Visits lost column shows only with Web Analytics. The Orders, revenue column shows only when Web Analytics has not recorded an order. If your shop records orders, see the orders and revenue lost in the outage panel and in the uptime card. The Bot errors column shows only with Server Log Monitoring. A dash means that the source has no figure for the outage.
The Kind chip
The Kind chip in the Every outage table of Uptime Monitoring names the failure. If the site sent an HTTP status code, the chip shows only that code, for example 503. A server error (5xx) shows in red. Without a code, the chip shows the kind. For Other, it shows the cause when the checker recorded one. The ? of the column lists the kinds:
| Chip | Meaning |
|---|---|
| 5xx | The server answered with an error. |
| 4xx | Blocked or not found, for example 403 or 404. |
| Timeout | No answer in time. |
| DNS | The address could not be found. |
| SSL/TLS | The certificate or the secure handshake failed. |
| Other | A network error or a refused connection. |
Hover the chip to see the cause in plain words, for example Connection refused, Redirect loop or Address not found (DNS).
Google back
The Google back column of the Every outage table shows how long crawling took to return to normal. The usual crawl is the average of the seven days before the outage. No drop means that crawling never fell to half. Without Server Log Monitoring, the column shows a dash.
Filter the outages
The filter bar of the Availability tab has filters for the Every outage list:
- Kind of failure: 5xx – server error, 4xx – blocked or not found, Timeout, DNS, SSL/TLS, Redirect loop, Could not connect or Other.
- Status code: the code that the site answered with when the outage began. No HTTP response means that the site sent no HTTP answer.
- Duration (seconds): more than or less than a duration in seconds. Use 60 to separate a short flap from a real outage.
- Error text: the error that the checker recorded, for example
connection refused.
The filter conditions narrow the outage count, the downtime, the recovery time and the outages list. The uptime, the chart and what the outages cost do not change. While a filter is on, the Uptime tab shows the time down in the matching outages, for example 12 min down in these outages. If no outage matches, the card shows No outage matches the filters.
One outage
Click a row in the Every outage table to open the outage panel. The title says Down for and the duration, or Down right now. Use the arrows to go to the previous or next outage. The panel has up to four tabs.
What happened
The What happened tab of the outage panel has three sections:
- What the checker saw: the raw error text of the checker.
- Response time around the outage: the answer time of the checks in the hour before, during and half an hour after the outage. A gap is a stretch where no check succeeded.
- Alerts that fired: the alert rules that sent a message for this outage. If none fired, click Set up an alert to create a rule in a side panel.
What it cost
The What it cost tab of the outage panel shows Visits lost, Orders lost and Revenue lost for this outage, from Web Analytics. Each figure is an estimate. Without Web Analytics, the tab offers Connect Web Analytics.
Search and AI bots
The Search and AI bots tab of the outage panel shows only with Server Log Monitoring. It lists the bots that got errors (Bot, Errors hit, Pages), and the pages that they got errors on (Page, Bot, Errors hit, Last status). If the logs do not cover the outage, the tab says so.
Before and after
The Before and after tab of the outage panel has two sections:
- Afterwards: Google back after, Pages Google got errors on (from Search Console URL Inspection) and Clicks, a week after (Google clicks in the 7 days after, against the 7 days before). These are next to the outage in time, but they do not prove that the outage caused them. Outages under 30 minutes are not looked at.
- Before the outage: what Page Monitoring saw change on your watched pages in the hour before. This can show a release or a change that broke the site. Without watched pages, it offers Watch a page.
Up, but something was wrong
The Hours when something was wrong although the site looked up card of Uptime Monitoring finds problems that a simple check cannot see. Our check asks for one page. Google asks for many pages. The card has two tabs.
The Google tab uses Server Log Monitoring. It lists the hours when verified Googlebot got errors, although every check passed. An hour shows when it has at least 20 Googlebot requests and at least 5 errors that are 10% or more of them. Errors are server errors (5xx) and too many requests (429).
The columns are Hour, What failed, Errors, Of its requests and Mostly in (the folder with most errors). Click an hour to see the pages that took the errors.
Google Search Console
The Google Search Console tab lists pages whose last fetch by Google got a server error, outside every outage of our checks. The data comes from URL Inspection. Google inspects a sample of your pages, so there can be more.
If a source is not connected, its tab shows Not connected and a button to connect it.
Connect more sources
The Availability tab of Uptime Monitoring can show a short line for each source that is not connected. The lines show only if the date range has outages:
- See how many visits each outage cost: click Install the Web Analytics script to add the visits, orders and revenue that each outage cost.
- See whether Googlebot hit your outages: click Connect server logs to see which search and AI bots got errors while the site was down.
After you connect a source, the line says that the data comes with the first data of that source.
Troubleshooting
Why does the Availability tab show No outages in this window?
If the Every outage card shows No outages in this window, your site answered every check in the date range. This is good. To see older outages, select a longer date range.
Why do Visits lost and Revenue lost show a dash?
If Visits lost shows a dash, Web Analytics has not priced the outage yet, or it is not installed. If Orders lost or Revenue lost shows Not tracked yet, set up conversion tracking. See Ecommerce.
Why does the Every outage card show Outages can't be counted while your server refuses our checks?
Your server or firewall turns our checks away. VitalSentinel then cannot tell if the site is down. Allow our monitoring in your firewall or CDN. See Monitoring IP addresses.
Why did Google get errors when my site was up?
Our check asks for one page. Google asks for many pages, so it can find errors on pages that our check never sees. Open the hour in the Google tab to see the pages that took the errors.
Related
- Uptime Monitoring - how the checks work and the page notices
- Response time - how fast your site answers
- Server Log Monitoring - what bots get from your server
- Ecommerce - conversion tracking for orders and revenue
- Alerts - get a message when your site goes down
- Monitoring IP addresses - allow our checks in your firewall
Overview
Uptime Monitoring checks all day that your site answers, records every outage and its cost, tracks response times, and watches your SSL certificate and domain.
Response time
The Uptime Monitoring Response time tab shows how fast your site answered our checks, the slowest checks, and how fast it is for visitors and Googlebot.
