A backup you have never restored is a hope
Backups get verified against their checksums, and restores get rehearsed. The time to discover an archive is incomplete is not the morning you need it.
06 · Service
Hosting & infrastructure
Plenty of agencies will build you something and hand you a zip file. The interesting question is who you call at 2am, and whether the person who answers has ever seen the inside of the thing that broke.
What this usually includes
- Containerised deployment with automated TLS certificates
- Metrics, log aggregation and dashboards, so an incident has evidence
- Backups — and, more importantly, restores that have actually been tested
- Uptime monitoring and alerting that reaches a human
- Deploys that can be rolled back
- Taking over hosting for something already built, after an audit
How we approach it
Rollback before launch, not after the first outage
Every deploy should have a way back. Set up beforehand it costs nothing; improvised during an incident it costs the incident.
Boring, documented, and not a hostage situation
Standard tools, written down, on infrastructure in your name where possible. Leaving should be an inconvenience, not a migration project.
What you end up with
- The running infrastructure, documented
- Dashboards and alerting you can see for yourself
- A written runbook for the things most likely to go wrong
- Credentials and ownership in your name
What this isn’t
- A formal uptime SLA with financial penalties — ask and we will talk about what is realistic, but we will not put a number on a page
- On-call cover for software we did not build and have not audited
Is this the thing you need building?
Describe the problem in a couple of sentences. You’ll get a straight answer about whether we’re the right people — including when we’re not.
[email protected]