CI and CD pipelines
Automated build, test and deploy pipelines on GitHub Actions, GitLab CI or your existing runner. Every merge produces an artefact you can trace, and every deploy has a rollback path that someone has actually tested.
Services · DevOps & Cloud Operations
DevOps and cloud operations sized for small teams
Pipelines that ship, infrastructure you can rebuild from code, and alerts that only wake someone when they should. We set up the operational layer your product needs, then hand it over with the runbooks to match.
You own the code and IP Weekly demos Reply within one business day
infra · ci/cd
$oldevs deploy --env production
✓Build passed · 212 tests · 0 warnings
✓Migrations applied · postgres 16
✓API v2 healthy · 142ms p95
→Rolling out to 3 regions…
API
Postgres
Redis
K8s
DevOps & Cloud Operations, in short
OlDevs is a full-stack technology studio in Vancouver, British Columbia that sets up and runs DevOps and cloud operations for small product teams.The work covers CI and CD pipelines, environment and branching strategy, infrastructure as code with Terraform, containers and orchestration, secrets management, and observability across logs, metrics and traces. We also put in place alerting and on-call basics, backups proven by restore drills, a written disaster recovery plan, a security patching cadence and ongoing cost control. Everything is right sized for a handful of engineers rather than an enterprise platform team, and clients own all code, accounts and IP.
Key facts
- 01Service
- DevOps and cloud operations for small teams
- 02Studio
- OlDevs, Vancouver, British Columbia, since 2014
- 03Core scope
- CI and CD, infrastructure as code, observability, backups, cost control
- 04Tooling
- Terraform, containers and managed cloud services, current at the time of writing
- 05Working style
- One accountable team and a working demo every week
- 06Ownership
- Clients own all code, configuration, accounts and IP
Operations sized to the team you have
What we set up and run
Infrastructure as code
Your cloud footprint described in Terraform and kept in version control, so environments can be reviewed, rebuilt and torn down on purpose. State is stored remotely with locking, and modules stay small enough to read in one sitting.
Containers and orchestration
Applications packaged as containers with sensible base images, health checks and resource limits. We run them on managed services such as Kubernetes, ECS or Cloud Run, and we will tell you when a container is more machinery than your team needs.
Observability and alerting
Logs, metrics and traces collected in one place with dashboards that answer real questions. Alerts are tied to symptoms users feel, routed to an on-call rota you can actually staff, and reviewed so the noisy ones get deleted rather than muted.
Secrets, access and patching
Credentials moved out of repositories and into a managed secret store, with least privilege roles and short-lived tokens. Dependency and image updates run on a set cadence so security patching is a routine chore instead of an emergency.
Backups, recovery and cost
Backups that are proven by scheduled restore drills, a written disaster recovery plan with a recovery point and time you have agreed to, and a monthly look at spend to right size instances, storage tiers and forgotten test environments.
Process
How an engagement runs
Discovery
A short review of your repositories, cloud accounts, deploy steps and incident history. We map what exists today and where the risk actually sits, then agree what good looks like for your team.
Plan
A written plan with a sequence, not a wish list: what gets automated first, what gets moved under Terraform, what stays as it is for now, and what each stage changes for the people doing the work.
Build
Pipelines, environments, modules, dashboards and secret storage built in small increments with a working demo every week. Changes go in behind review, with a rollback path for each one.
Prove
A restore drill, a failover test and a deliberate look at alert quality before anyone calls it finished. We fix what the rehearsal exposes and record the recovery times we measured.
Handover
Runbooks, architecture notes and a live walkthrough with your engineers, plus a support window while the rota settles. Request a quote to start the conversation.
How we work
How we approach the work
01
Start from your constraints
Team size, budget, compliance rules and who carries the pager decide the architecture. We design for the organisation you have now, with room to grow, not for a platform team you have no plans to hire.
02
Automate the path you repeat
The build, test and deploy sequence gets automated first, because that is where manual steps cause the most damage. Once a deploy is a single reviewed merge, everything downstream gets calmer.
03
Describe infrastructure in code
Networks, databases, queues and permissions live in Terraform under review, so a change is a pull request rather than a console click nobody witnessed. Environments become reproducible instead of precious.
04
Make the system observable
Structured logs, useful metrics and traces across service boundaries, with dashboards built around the questions you ask during an incident. If a dashboard never gets opened, we remove it.
05
Rehearse the bad day
Restore drills, failover tests and a disaster recovery plan written in plain language. We measure how long recovery really takes and fix the slow parts before an outage does the measuring for us.
06
Hand over properly
Runbooks, architecture notes, access in your own accounts and a walkthrough with your engineers. The goal is a system your team can operate without us, with clients owning all code, designs, accounts and IP.
Who it's for
DevOps & Cloud Operations for organisations that have to get it right.
Corporations
Internal product groups that need release discipline, audit friendly change history and environments that match, without waiting on a central platform team for every request.
Associations and government
Public sector and member organisations with procurement rules, data residency requirements and accessibility obligations. We document decisions clearly and keep every account in your own name.
Franchises
Multi location brands running one platform across many sites. Shared pipelines, per region environments and clear rollout controls so a change reaches one location before it reaches all of them.
Entrepreneurs and startups
Founders who need to ship quickly without building operational debt they will pay for later. We set up the smallest sensible stack and leave you able to run it yourselves.
Weekly
Working demo cadence
2014
Studio founded
1 day
Reply to every enquiry
FAQ
DevOps & Cloud Operations — questions we hear first.
Usually not. Kubernetes earns its place when you have many services, several teams and real scaling pressure. For most small teams a managed container service or a well configured virtual machine group is cheaper to run and far easier to reason about. We recommend the smallest setup that meets your reliability and compliance needs, and we say so plainly.
We work with the major providers, including Amazon Web Services, Google Cloud and Microsoft Azure, and with smaller hosts where they fit. The choice usually comes down to what your team already knows, what your data residency rules require and which managed services remove the most work. We are not a partner or reseller for any vendor, so the recommendation stays yours.
Yes. Most engagements start with infrastructure that grew by hand: a few servers, some console clicks and a deploy script only one person understands. We document what exists, bring it under Terraform in stages, and avoid a rebuild unless the current setup is genuinely blocking you. Nothing is changed without a rollback path.
We help you build an on-call practice rather than replace it: rotas, severity levels, escalation paths and a short runbook per alert. For a defined period after launch we can share the rota so your team learns the system with support beside them. Long term ownership stays with you, which is what keeps the knowledge in house.
By making spend visible before it becomes a surprise. That means tagging resources by environment and owner, setting budget alerts, right sizing instances after watching real usage, moving cold data to cheaper storage tiers and switching off environments nobody uses. We review it on a regular cadence and show you the reasoning behind every change.
On a schedule we agree at the start, and always after a significant change to the data layer. A backup nobody has restored is a hope, not a plan, so drills are run against a separate environment and timed, then written up with the recovery point and recovery time we actually achieved.
Everything. Clients own all code, configuration, designs, cloud accounts and IP, including the Terraform modules, pipeline definitions, dashboards and runbooks we write. Access is set up inside your own organisation from day one, so there is no lock in and no waiting on us to hand over a login. Request a quote to talk through scope.
Let’s connect
Let’s scope your devops & cloud operations project.
Tell us what you’re building. We’ll reply within one business day with next steps and a tailored quote — no obligation.