Service status

What we monitor, how an incident reaches a person, and where it gets written down. A live feed is being built; until it exists this page says so.

What is monitored

Five things, on every service we operate, whether or not the client asked for them.

Client websites and applications. Reachability from more than one location, response time against an agreed threshold, and application errors.
Cloud and dedicated servers. Load, memory and disk with the trend, so a full disk is a ticket rather than an incident.
Business email. Delivery, authentication records and queue depth. Most mail problems are silent until somebody says a message never arrived.
Certificates. Watched well before expiry rather than on the morning of.
Backups. Verified by restoring them, not by reading the job's own report of itself.

How an alert becomes a person, and the response times written into the agreement rather than implied by a badge: monitoring and response.

An incident on a service we operate is reported to the client directly, by a person. If you are seeing one before we are, tell us.