Cloud & infrastructure
We run our own. That’s why we can run yours.
Most software companies our size deploy to a managed platform and call it DevOps. That is often the right choice — but it is not the same as operating infrastructure, and when something breaks at 2am the difference is the only thing that matters.
Our own stack
This is what sits behind the product we sell.
Not a capability list. This is the infrastructure Optician Dynamics runs on, operated by the same people who would operate yours.
Self-hosted source & CI
Our own GitLab and our own runners, not a hosted plan. Pipelines, artefacts and secrets stay on infrastructure we control.
More than one provider
Workloads across Hetzner, Contabo and AWS. Multi-provider by choice — it keeps costs honest and avoids a single vendor holding the business.
Containers across VMs
Docker across multiple machines, with Nginx Proxy Manager at the edge and Cloudflare in front of it.
Our own observability
A self-hosted SigNoz stack — traces, logs and metrics. Alerts fire on the paths that carry money, not just on whether a box is up.
Per-tenant migrations
Schema changes fanned out across every tenant database automatically, with drift detection when one falls behind.
Deploys that verify themselves
Post-deploy checks that fail the pipeline if the live build is not the commit that was built. Green is meant to mean something.
The part nobody advertises
Migrations across a fleet of live databases.
Shipping a schema change to one database is easy. Shipping it safely to every tenant, in order, and knowing which ones lagged behind, is the work — and it is where multi-tenant systems quietly rot.
We built the fan-out, the drift detection and the verification because we needed them ourselves. Anyone can describe this. Far fewer have had to fix it at 2am.
- Migrations fanned out across every tenant, in order
- Drift detection when a database falls behind the fleet
- Rollback paths that have been exercised, not just written down
- Alerting on the transactions that matter, not on CPU
- Backups that have actually been restored from
- A build that refuses to pass when the deployed commit is wrong
What we do for you
Move it, run it, and answer the phone.
Cloud migration
On-premise or one provider to another, incrementally, with the old system running until the new one has proved itself.
Managed hosting
We host and operate it: patching, certificates, backups, capacity. You get an environment and someone accountable for it.
Monitoring & on-call
Instrumentation on the paths that carry money, alerting that reaches a human, and a rota behind it.
CI/CD
Pipelines to staging and production with gates that mean something, so a bad deploy fails loudly instead of quietly.
Backup & recovery
Backups tested by restoring them. An untested backup is a hope, not a plan.
Cost review
Most bills contain something nobody has looked at in a year. We will tell you where yours is going, including when the answer is "nowhere useful".
Use a managed platform when…
- One application, one database, modest traffic
- No compliance or data-residency requirement pushing you elsewhere
- Your team is comfortable owning the platform themselves
- Cost of the platform is not yet a meaningful line in the budget
We mean it. If that is you, a good PaaS will serve you better and cheaper than we would, and we would rather say so now.
Call us when…
- Multi-tenant, where one migration touches many databases
- Downtime costs money the same hour it happens
- Data residency or regulation constrains where it can run
- Nobody internally wants to own the pager
- Your cloud bill has grown and nobody can explain why
Who runs it after we hand it over?
The honest answer, from most vendors, is you. Ours is that we can — monitoring, backups, migrations, upgrades and someone to call, on a monthly arrangement.