Get in touch
≡

Reliability & operations

Built, shipped, and kept running.

Most agencies build an application and hand it over. We host it, monitor it and maintain it on infrastructure we own and operate. Managed hosting, application support and disaster recovery are part of the engagement, not an upsell afterwards.

The question this answers

"What happens in six months when it breaks and you've moved on?"

That is the buyer's real fear, and it is a fair one. Everything below is running in production today across our own products and our client estate. It is written as specifics rather than reassurance, because specifics are the only part of a reliability claim that can be checked.

The register

What is in place, and what it actually means.

ControlPolicyState
InfrastructureOwned and provisioned by us — a single well-provisioned server running the full estate, with room to grow. Not resold hosting, so there is no third party between a fault and the person who can fix it.Operated
TLS certificatesAutomatic for every domain, including per-customer custom domains provisioned on demand. A tenant points their domain at the platform and gets HTTPS with no manual step and no ticket.Automated
BackupsNightly, offsite, with tiered retention — daily, weekly and monthly generations kept to a written policy rather than growing forever.Nightly
Restore testingBackups are pulled back down from offsite storage on a schedule, restored into a scratch database, and row counts compared against production. A backup that has never been restored is a hypothesis, not a backup.Scheduled
MonitoringFails loudly. A failed or stale backup raises an alarm on sight; it cannot go unnoticed for weeks, which is the normal failure mode of a backup system nobody watches.Alarming
MigrationsZero downtime. The entire estate was recently moved between providers with no service interruption, verified afterwards by a cold reboot of the new host.Verified
Application maintenanceDependency and platform updates, mobile OS and store-policy changes, certificate and domain renewals, and the small breakages that follow other people's releases.Continuous

Every row is a fact, not a promise. If a row cannot be stated as a specific, it does not go on this page.

Disaster recovery

What happens when something does break.

  1. The alarm fires before you notice

    Monitoring watches the application, the host and the backup pipeline itself. Silence is not treated as success — a check that stops reporting is a failure.

  2. An engineer looks, not a queue

    The person reading the alert is the person who wrote the software. There is no first-line triage layer to escalate through.

  3. Restore from a backup we have already restored before

    Because restores are tested on a schedule, recovery is a procedure that has been rehearsed rather than a procedure being attempted for the first time under pressure.

  4. You get told what happened

    A written account of the fault, the fix, and the change that stops it recurring. Not a status-page colour change.

Service levels

What an agreement looks like.

Indicative service tiers. Final terms are written into the engagement — ask us for the current SLA document.
IncludedStandardExtended
Hosting on owned infrastructureYesYes
Automatic TLS, all domainsYesYes, incl. customer custom domains
Backup frequency / retentionNightly · daily, weekly, monthlyNightly · extended monthly retention
Restore testingScheduled, row-count verifiedScheduled, with a written restore report
Monitoring & alertingApplication, host, backup pipelinePlus business-level checks you define
Application maintenance windowBusiness hours, both officesAgreed response target, out of hours
Incident reportWritten, per incidentWritten, plus a quarterly review

Proof

The estate we operate is our own.

Theourgia, AstroPractise, RebetiChord and Kotobaseed all run on this infrastructure — including Kotobaseed's per-tenant custom domains and live video. We are the first people affected by our own operational decisions.

See the productsAsk about an SLA

Next step

Already have an application nobody is maintaining?

We take over existing estates: read the code, document it, put backups and monitoring in place, then keep it running.

Talk about a takeoverSee our work