Uptime reports for clients, what to include every month

Most clients judge a maintenance retainer by what they notice. When everything works, they notice nothing, and the retainer starts to look like a cost with no return. A short, regular report changes that. It shows that the site was watched every minute, that problems were caught and handled, and that expiring certificates and domains were dealt with before they became emergencies.

The report also protects the relationship when something goes wrong. A client who has received calm, accurate reports for a year reads an incident summary as one line in a long record of reliable work, not as a surprise.

Clients rarely care about response time percentiles for their own sake. They want answers to four questions. Was my site available? If not, for how long and why? Is anything about to expire or break? What did you do for me this month? Build the report around those questions and put the technical detail below them.

• Uptime percent for the month, per site or per key page such as the homepage and the checkout
• The number of incidents, each with start time, end time, duration and a one line cause
• Response time for the month as a median and a 95th percentile, with the trend against the previous month
• Days left on the SSL certificate and the domain registration for every hostname
• Maintenance windows that took place and whether they were excluded from the uptime figure
• A short summary of the work done, written in plain language
• What to watch next month, such as a renewal or a planned migration

Put a summary at the top: the uptime figure, the number of incidents, and one or two sentences on the month. Everything else goes on the following pages. A client who reads only the first page should still come away with an accurate picture.

Uptime percent is the share of the month during which the site was confirmed available. A month of 30 days has 43,200 minutes, so 99.9 percent allows about 43 minutes of downtime and 99.99 percent about 4 minutes. For the full math, see 99.9 uptime explained.

Three choices decide whether the number is fair, and they should be stated in the report so nobody argues about them later.

1. 01 What counts as down. A failure confirmed from a second region is a fair definition. A single failed check from one location is not, because it often reflects the network between the probe and the site rather than the site itself.
2. 02 What is excluded. Planned maintenance announced in advance is usually excluded, and the report should list those windows so the exclusion is visible.
3. 03 What is measured. The homepage is a minimum. For a store, the product page and the checkout matter more, and for a lead generation site, the form page.

Report the number you measured, even when it is lower than you would like. A report that is always 100 percent stops being believed.

Write each incident in four short parts: what happened, how long it lasted, what caused it, and what was done so it does not happen again. Avoid jargon, and avoid blaming the client or another vendor in the report itself.

• What happened. The checkout page returned errors and orders could not be placed.
• How long. From 14:05 to 14:31, 26 minutes.
• Why. A plugin update conflicted with the payment step.
• What we did. We rolled back the update, tested the checkout and scheduled the update again after a fix from the plugin author.

If the client already received live updates during the incident, through an alert or a status page, the report confirms the final duration and the cause.

The report is part of your service, so it should carry your agency logo and colors, not the name of a tool. White-label reports keep the attention on your work and avoid the question of why the client should pay you for a tool they could buy directly.

Send the report on the same day every month, for example on the third working day, so clients learn to expect it. Email the PDF to the contact who approves the retainer, and keep the same structure month after month so changes stand out.

When an agency runs many client sites, the report should be built per client, with only that client's sites in it. Organize monitors into one group per client from the start, so reports, alerts and status pages all follow the same structure. If your current tool makes that hard, our UptimeRobot alternative for agencies comparison shows what changes when groups, seats and reports come in one plan.

1. 01 Summary: uptime for the month, incidents count, one sentence on the month.
2. 02 Availability: uptime per site or key page, with the previous month for comparison.
3. 03 Incidents: each with times, duration, cause and action taken.
4. 04 Performance: median and 95th percentile response time, with the trend.
5. 05 Renewals: certificate and domain expiry dates for every hostname.
6. 06 Maintenance: planned windows and how they were counted.
7. 07 Work done: updates, fixes and improvements in plain language.
8. 08 Next month: renewals, planned changes and recommendations.

UptimePilot measures uptime with confirmation from a second region, stores every incident with its duration, and tracks certificate and domain expiry for every hostname. From Team upward you get a monthly SLA report as a PDF, and Business and Enterprise add white-label client reports with your logo and colors, built per client group. See uptime monitoring service reports and uptime monitoring for agencies.