Status

Current status

Nothing updates this page automatically. If Kromdraw isn't answering, email me and you get a person rather than a ticket number.

Monitoring in place

The database is backed up continuously: a base backup every night and every write archived on top of it. A daily job checks that the archiving is still advancing and fails loudly in the job queue when it is not. Routine cleanup removes expired sessions and temporary files according to the privacy policy.

What is not in place

• No external uptime monitor, so an outage is noticed when somebody notices it.
• No error-tracking service. Errors go to the log on the machine.
• No status feed to subscribe to, and no automatic email when something breaks.
• No published uptime figure, since nothing measures one.
• Nothing sends an alert. The daily backup check fails into the job queue, where it waits until I look at it.
• The restore drill is run by hand rather than on a schedule, so what proves the backups work is me running it.

How incidents get communicated

I email the people affected as soon as I can give them useful, accurate information without making the incident worse. Depending on its severity, I also publish a report on this page. A retrospective follows as soon as it can be written safely and accurately. There is one person responsible for incidents, so response times depend on when an incident is noticed.

Where your data lives covers what is backed up and how often.