What does website uptime monitoring mean?
Scheduled automated checks establish whether a specified address responds and can trigger an alert on failure. They do not observe every second or every function. Over a 365-day year, 99.9% uptime corresponds to about 8.76 hours unavailable; this is arithmetic, not our guarantee.
What does the check actually do?
A probe sends a request and compares the response with a defined condition. That may be a successful HTTP status, specified text or completion of an action. A responding homepage does not establish that a form sends or a payment completes.
Catchpoint also discusses dependencies: DNS, the content delivery network (CDN), servers and outside services. Not every tool checks these automatically. Record which addresses and functions are monitored, from where, and what counts as failure.
How often do different tools check?
UptimeRobot lists five minutes for Free, marketed for hobby and non-profit projects. Published intervals are 60 seconds for Solo, 30 for Team and 15 for Scale; Enterprise offers custom intervals. Check the current plan and terms before choosing.
A shorter interval may reduce the wait for the next probe. It is not a notification or repair deadline. Repeated checks, confirmation and alert delivery take time too. A short problem between probes may go undetected.
What does the percentage mean?
For the same year, 99.99% corresponds to about 0.876 hours unavailable — under an hour. This is total time, not incident count: it could be one incident or many short ones. The percentage does not imply one outage each month.
Compare matching periods, scope and rules. A successful-probe percentage does not establish continuous availability for every visitor. A stricter target may suit the impact of interruption, but is not automatically required for every site taking orders.
What are false alarms?
A failed probe may reflect the route to the site rather than a general outage. Cloudways recommends checks from different locations and suitable thresholds. Extra confirmation may reduce noise but delay the alert.
Do not dismiss a regional problem merely because another location gets a response. Define failed-probe counts, recovery conditions and recipients. Test delivery; an event recorded in a dashboard does not establish that someone was notified.
What should be written down?
Separate check frequency, availability target and response deadline. If there is a service level agreement (SLA), define the period, exclusions, measurement and consequences of missing it. These depend on the agreement, not the word “monitoring”.
In reporting, distinguish an observed interval of failed checks from an exactly established outage duration. The problem may start or finish between probes. Other records may be needed to establish timing more precisely.
Questions
Does it catch a slow but working site?
Some tools measure response time or test particular actions. That needs configuring and describing. A successful response alone does not establish fast loading or a working order journey.
Does a brochure site need 99.99%?
Not necessarily. Consider interruption impact, required functions and budget. The polling interval is not a promised availability percentage.
Who sets the alert threshold?
You or the provider within the agreed scope. Consecutive failures and confirmation from other locations may limit noise but take time. Agree the hours when a person responds too.
Does a small site need a separate SLA contract?
Do not assume size alone requires a separate agreement. Written checks, responsibilities and deadlines help comparison with reporting. A contractual availability target and SLA service credits depend on the particular agreement.
Sources
Related reading
Next step
Record what is checked and when someone responds.
Our reachability check runs every hour under the written inventory. Detected failures and their observed intervals appear in the monthly report. This does not guarantee an availability percentage or detection of every short outage; duration is reported according to the available observations.
or write to contact@ergostrat.com