Response time monitoring

How fast does your site answer?

Every check AlertSleep makes also times the answer, from the DNS lookup to the last byte of the response, measured from our server in Poland. You see it as a chart, a 30-day average and a figure in your SLA report, and a site that stops answering in time counts as down.

  • Timed on every check, every 60 seconds
  • Charts from the last hour to 7 days
  • No answer within the timeout = down, and you are alerted

Response time, last 24 hours

example.com
avg · 182 ms
api.example.com
avg · 94 ms
shop.example.com
avg · 640 ms

What is measured

One number per check, in milliseconds. This is what goes into it.

  1. 1

    The request your monitor defines

    The same request as the uptime check: your URL, method, headers and body, sent from our server in Poland.

  2. 2

    The clock runs to the last byte

    Timing starts before the DNS lookup and stops when the whole response has arrived: connection, TLS handshake, the time your server takes to produce the answer, and the transfer of the HTML or JSON it returns.

  3. 3

    Stored with the check

    The figure is saved with the check result. Charts, averages and reports are built from those numbers; nothing is sampled or estimated.

  4. 4

    No answer in time is a failure

    If the full response has not arrived within the monitor's timeout (5 to 60 seconds, 30 by default), the check fails and the usual rules apply: the failure threshold, confirmation from a second location and your alerts.

Where you see it

The same measurements, wherever you look at a monitor.

Response time chart

Each monitor's page in the dashboard charts its response times for the last hour, 6 hours, 24 hours or 7 days.

30-day average

Every monitor shows its average response time over the last 30 days, from successful checks only, so an outage does not skew it.

In the SLA report

The SLA report lists each monitor's average response time for the period beside its availability, from daily figures that are kept after individual checks are deleted.

On your status page

A public status page shows visitors each monitor's average response time and a line of its response times over the last 24 hours.

History by plan

Individual checks are kept for 30 days on Studio and 90 days on Business.

What it does not measure

Response time monitoring answers one question: is my server quick to answer? It is not a browser.

  • No Core Web Vitals (LCP, INP, CLS) and no Lighthouse score: nothing is rendered, and images, CSS and JavaScript are not downloaded.
  • No breakdown into DNS, connection, TLS and time to first byte: you get the total time of the exchange.
  • No alert for a site that is slow but answers. The alert comes when a check fails, including when the response takes longer than the timeout.
  • Times are measured from our server in Poland. A visitor on another continent will see different times.
  • For page experience in real browsers, use Google PageSpeed Insights or the Core Web Vitals report in Search Console.

Frequently Asked Questions

What does response time measure?

The time from the start of a check to the last byte of the response: DNS lookup, connection, TLS handshake, the time your server spends producing the answer and the transfer of the response itself. It is measured on every check, every 60 seconds at the shortest interval.

Does AlertSleep measure Core Web Vitals?

No. Core Web Vitals (LCP, INP, CLS) are measured in a browser that renders the page and loads its images, CSS and JavaScript. AlertSleep sends an HTTP request and times the answer; it renders nothing. Google PageSpeed Insights and the Core Web Vitals report in Search Console measure page experience.

Will I be alerted when my site gets slow?

When it gets too slow to answer. Each monitor has a timeout between 5 and 60 seconds (30 by default). A response that has not fully arrived by then makes the check fail, and the failure is confirmed and alerted like any other. There is no separate alert for a site that is slow but still answers.

Where is response time measured from?

From our server in Poland. A site hosted far away shows the network distance in its times, so compare a monitor with its own history rather than with another site.

How long is the history kept?

Individual checks are kept for 30 days on Studio and 90 days on Business. Daily averages are kept for the SLA reports after the individual checks are deleted.

Is there a free way to test response time?

Yes: the free response time test times any URL once, with no account. Monitoring repeats that measurement on every check and keeps the history, from $59 a year after a 14-day trial.

Know when your site stops answering in time

Response time is recorded on every check of every HTTP and keyword monitor, from $59 a year. Try it free for 14 days.