How to check site downtime
HTTP Status Code Checker
Free, no sign-up. We only fetch the public URL you enter.
Short answer
To check site downtime right now, test the URL from an independent server, like the free check above: a 5xx status, a timeout or a connection error means the site is down. To find out when and how long a site was down in the past, you need records: the incident history of an uptime monitor, the hosting provider’s status page and incident reports, server access and error logs (gaps or 5xx bursts), or analytics showing a sudden drop to zero visits. Calculate uptime as available time divided by total time; 43 minutes of downtime in a month equals 99.9 %. monitordefend records every outage with start, end and duration, checking every five minutes.
- Now: run an external check and read the status code.
- Past: monitoring history, hosting status page, logs, analytics.
- Duration: from the first failed to the first successful check.
- Last successful monitor check
- Error log: database connection refused
- Monitor: two failed checks, incident opened
- Access log: first 200 responses again
- Monitor: check successful, incident closed
“Check site downtime” can mean two different things. Sometimes you want to know if a website is down right now. Sometimes you need to know whether it was down last night, how long the outage lasted, how often it happens or whether your host kept its uptime promise. The first question takes seconds; the second needs data that someone recorded while it happened. This page covers both: how to check the current status with the free tool above, where to find information about past downtime, how to calculate downtime and uptime, and how to make sure you have the records next time.
Check if a site is down right now
Enter the address in the free check above. It requests the page from a server in Germany, follows redirects and shows the status code, final URL, response time, server and IP address. A status between 200 and 399 means the site is up. A 5xx status, a timeout or a connection error means the site is down for everyone, not just for you.
If the check is fine but the site does not load in your browser, the problem is local: DNS cache, browser extensions, VPN, firewall or your internet connection. If both fail, the server, the hosting or the domain configuration is the problem.
Check past downtime with an uptime monitor
The most reliable record of past downtime comes from an uptime monitor that was already running. It checked the site every few minutes and logged each failure, so its incident list tells you when each outage started, when it ended, how long it lasted and which error occurred. Most monitors also show daily uptime bars and percentages for the last weeks or months.
Keep in mind the resolution: with a five-minute interval, start and end times are accurate to a few minutes, and very short outages between two checks are not recorded. For past periods without monitoring, the following sources help.
Check your hosting provider’s status page
Most hosting providers, cloud platforms and CDNs publish a status page with current incidents, past incidents and planned maintenance. If your site was down because of a platform problem, it is usually listed there with times and an explanation. Platform status pages do not show problems specific to your server or application, so a quiet status page does not prove your site was up.
Check server logs
Web server access logs record every request with time, URL, status code and often the response time. During an outage you typically see either a burst of 5xx codes or a gap with no entries at all, if the server was unreachable. Error logs and application logs often name the cause, for example a database connection error, a memory limit or a crashed process. Many hosts make logs available in their control panel for a limited number of days.
22:42:51 GET / 200 0.41 s 22:43:07 GET /shop 502 30.00 s 22:43:12 GET / 502 30.00 s ... 42 minutes of 502 responses 23:28:03 GET / 200 0.56 s
- Access log: look for 5xx bursts and gaps in requests
- Error log: look for fatal errors, timeouts, memory limits
- System logs: restarts, out-of-memory kills, disk full
- Deployment history: changes shortly before the outage
Check analytics and search data
If you have no monitoring and no logs, analytics can give a rough picture: a sudden drop of visits to zero for a period, followed by a return to normal, usually means downtime. Hourly views make this visible. Google Search Console’s crawl stats report shows server errors and failed requests by Googlebot, which also hint at availability problems. Both are indirect and imprecise, but better than nothing.
Check downtime of other websites
For third-party services, look at their official status pages, which often include an incident history. Public outage report sites collect user reports and show spikes when many people have problems at the same time. Social media searches for the service name plus “down” show what users experience. For your own suppliers and critical services, consider monitoring their public endpoints yourself.
How to calculate downtime and uptime
Downtime duration is the time from the first failed check, or the first error in the logs, to the first successful check after the outage. Uptime percentage is the available time divided by the total time in the period. A month of 30 days has 43,200 minutes; 43 minutes of downtime equal 99.9 % uptime, 432 minutes equal 99 %.
Agree on what counts as downtime: only full outages, or also errors on important pages, or responses slower than a threshold? Service level agreements define this precisely and often exclude planned maintenance. The free uptime calculator on this site converts between percentages and minutes.
Checking downtime against a hosting SLA
Many hosts promise 99.9 % uptime and offer credits if they miss it. To claim them, you usually need your own evidence: dates, times, durations and error types, often within a short deadline. Monitoring data from an independent service is the strongest evidence. Read the SLA for exclusions, such as maintenance windows, problems caused by your own software or outages shorter than a few minutes.
Partial downtime and slow periods
Not every outage is total. Sometimes only the checkout fails, only one region cannot connect, or the site answers but takes 20 seconds. Visitors experience this as downtime even if a homepage check stays green. When you check past downtime, look at errors on important URLs and at response times as well. Monitoring several key pages and setting a response time threshold turns these partial outages into recorded incidents instead of vague complaints.
Writing a short incident report
For customers, managers or your hosting provider, summarise each significant outage in a few lines: start and end time with time zone, duration, affected pages or services, visible symptom, cause, fix and what will prevent a repeat. Attach the monitoring data or log excerpts. Collected over a year, these reports show patterns that justify better hosting or changes to the update process.
Finding the cause
Once you know when the downtime happened, look at what happened at that time: deployments, plugin updates, cron jobs, backups, traffic peaks, DNS or certificate changes, provider incidents. Recurring outages at the same time of day often point to scheduled tasks. Gradually rising response times before the outage point to resource exhaustion.
Make sure you have the data next time
Set up an uptime monitor now, keep server logs for at least a few weeks, subscribe to your hosting provider’s status notifications and write a short note after each incident with times, cause and fix. Then the next time someone asks “was the site down yesterday?”, the answer takes seconds.
Downtime records with monitordefend
monitordefend checks each website every five minutes, confirms failures with a second attempt and records each outage as an incident with start, end, duration and error. Uptime and response time are shown per day, week, month and 90 days, and you get an email when an outage starts. Its other checkers watch links, indexing, traffic, checkout and security signals. It does not offer public status pages or SLA reports.
Frequently asked questions
How can I check if a site is down right now?
Use the free check above. It tests the site from an independent server and shows the status code and response time.
How can I find out when my website was down?
Look at the incident history of your uptime monitor, your hosting provider’s status page, server access and error logs, and hourly analytics data.
How do I calculate downtime percentage?
Divide the downtime by the total time of the period and multiply by 100. Uptime is 100 % minus that value.
Can I see past downtime of a site I do not own?
Only indirectly: through the operator’s status page, public outage reports or your own monitoring if you had set it up.
How much downtime is normal?
Many business websites reach 99.9 % or better, which is about 43 minutes per month. Frequent or long outages point to hosting or application problems.
How precise is the downtime duration from a monitor?
It is accurate to about one check interval. With five-minute checks, start and end times can be off by a few minutes, and very short outages may not be recorded.
Why does my monitor show downtime my host does not confirm?
Your site may fail for reasons outside the host’s platform, such as application errors, or the host may exclude short outages and maintenance.
Related pages
Record every outage automatically
Start for free and set up your first check in minutes.
Start for free