Skip to main content

DevOps as a Service

Your infrastructure, on autopilot

We build and run the pipelines, clusters, and on-call rotation behind your product, so your engineers ship features instead of fighting YAML.

No lock-in. Everything we build lives in your repositories.

A deployment pipeline: commit, build, test, scan, deploy. 340 tests pass, no critical vulnerabilities, deployed to production in 8 seconds.

Built on the tools your team already knows

  • OpenStack
  • Ceph
  • Docker
  • Kubernetes
  • Terraform
  • Ansible
  • Prometheus
  • Grafana
  • GitLab
  • Vault
  • Helm
  • Argo CD

What we put in writing

Every commitment below is repeated in the contract, the pricing page, or the monthly report. A promise that lives only in marketing is not a promise.

  • A contractual SLA

    On Enterprise, missing the agreed response time carries a financial penalty, credited against the following month's invoice.

  • 24/7 on-call

    Our engineers join the rotation and answer the pager. Fifteen-minute response on Growth, five minutes on Enterprise.

  • Your repositories, your code

    Everything we build is committed to your version control under your licence. No proprietary agent, no hidden control plane, no exit fee.

  • Measured, then reported

    SLOs are agreed before we start and reported every month, with a blameless post-mortem after every incident, whether or not it flatters us.

What changes in the first ninety days

Before

Releases happen at midnight, by one person

With MizbanDevops

Any engineer ships during business hours

Before

Customers report the outage first

With MizbanDevops

Alerts fire before users are affected

Before

Nobody dares upgrade the cluster

With MizbanDevops

Upgrades are scheduled and uneventful

Before

Security is remembered before the audit

With MizbanDevops

Every merge request passes a security gate

Boring technology, chosen deliberately

We pick the tool your team can hire for and debug at three in the morning. Novelty is a cost paid by whoever is on-call, and that is usually you.

Source & delivery

  • GitLab CI
  • GitHub Actions
  • Argo CD
  • Helm

Runtime

  • Kubernetes
  • Docker
  • OpenStack
  • Ceph

Observability

  • Prometheus
  • Grafana
  • Loki
  • OpenTelemetry

Security

  • Vault
  • Trivy
  • OPA
  • Gatekeeper

Nothing here is proprietary to us. Every component is open source, widely deployed, and documented well enough that another team could take over tomorrow.

A pipeline, applied to ourselves

Every engagement runs the same four stages. You always know which one you are in.

  1. 1

    Week 1

    Assess

    We audit your repositories, infrastructure, and incident history, then produce a written findings report with a prioritised backlog.

  2. 2

    Weeks 2–6

    Automate

    Pipelines, infrastructure as code, and environments. We work in your repositories, in merge requests your team reviews.

  3. 3

    Weeks 6–10

    Observe

    Metrics, logs, traces, dashboards, and SLO-based alerting. We tune until the alerts are trustworthy.

  4. 4

    Ongoing

    Operate

    We join the on-call rotation, hold the SLOs, run blameless post-mortems, and report monthly.

What our clients measure

  • 14×

    Deployment frequency

    from fortnightly to multiple times a day

  • 23m

    Median time to recovery

    down from just over four hours

  • 99.98%

    Measured availability

    across managed production services

  • 31%

    Cloud spend reduction

    after right-sizing and idle reclamation

Median change across managed accounts in their first six months.

A monthly retainer, or a fixed-scope project

Managed service is priced per month by the size of your estate. Migrations and audits are scoped and quoted separately.

Starter

890USDper month

Small teams with a single production environment

Growth

Most chosen

1990USDper month

Scaling teams running multiple environments

Enterprise

Custom

Organisations with compliance requirements or on-premise estates

We went from dreading deployments to not thinking about them. The first month they mostly listened and read our code. The report they produced was more honest about our infrastructure than we had been with ourselves.
Head of EngineeringFintech platform, Tehran

Start with the assessment

One week, a written report, and a prioritised backlog. If you take nothing further, the report is still yours to keep.