Monitoring
Uptime monitoring for web agencies: what to check, how often, and what should happen next
Last checked
The worst way to learn a client's site is down is a text from the client. Monitoring exists to reverse that order: you know first, you act, and the client hears about it from you, if at all. For an agency the hard part is not the ping. It is what happens after it fails, across 20 or 40 sites, at 2am.
What to monitor on every client site
| Check | What it catches |
|---|---|
| Uptime | The site is not answering, or is answering with an error. |
| SSL certificate | The certificate is about to expire. Browsers then show a full-page warning, and to the client that is the site being down. |
| Domain expiry | The registration lapses. This one is rarely monitored and always catastrophic. |
| Performance | The site is up but slow enough to cost the client enquiries. |
| Homepage change | The page content changed when nobody on your team changed it: a hack, a bad deploy, a rogue plugin. |
These are the five checks Bluprent runs on every site you add, on a schedule you set.
How often
The default check interval in Bluprent is every 5 minutes. You can set it anywhere from every 60 seconds to once a day per site. A shorter interval finds an outage sooner and produces more noise from one-off blips; a five-minute interval suits most brochure sites, and a shop or booking site earns a tighter one. Domain and SSL expiry do not need minutes. A daily look is plenty.
What should happen when a check fails
A monitor that only sends an email leaves the actual work to you. A useful one runs a lifecycle:
- An incident opens by itself when a check fails, with the start time recorded.
- You are told, and so is the client, by email, without you writing anything.
- It closes by itself the moment the site recovers, so the duration is exact and nobody has to remember to clear it.
- If it stays open too long, it becomes a task. An outage that does not recover on its own needs a person, and it lands on your board.
Every incident then sits in the client's record, which is what turns "the site was down for two hours in March" from an argument into a line in the monthly report.
Why this belongs next to the CRM
A standalone monitor knows a URL. It does not know whose site it is, what plan they are on, or whether their invoice is overdue. When monitoring lives in the same record as the client, an incident can notify the right person, show up in the client portal and feed the uptime figure in the report. That is the argument for a CRM for website maintenance over a separate monitor.
Monitoring is one line of a care plan. It is included on every Bluprent plan, including the free one for up to three websites. See pricing.