Skip to main content
    Reign · Deployment

    Two models are available. Two are in development. One will not exist.

    Reign runs as a single-tenant dedicated instance, in our environment or in the customer’s own cloud account on AWS or Azure. Both are available today and both are single-tenant. Air-gapped and sovereign are in development. This page states which is which.

    If a deployment model is not listed as available below, treat it as not available.

    Deployment models

    A single-tenant dedicated instance, in our environment or in yours.

    ModelStatusWhat it is
    Single-tenant dedicated instanceAvailable todayThe instance is provisioned for one customer and serves that customer only.
    Customer cloudAvailable todayDeployment into the customer’s own cloud account on AWS or Azure. Google Cloud is planned for 2027. The instance is still provisioned for one customer and serves that customer only.
    Air-gappedIn developmentAir-gapped deployment. Not available today, and not being sold.
    SovereignIn developmentSovereign deployment. Not available today, and not being sold.
    Multi-tenant or shared SaaSNot on offerThere is no multi-tenant option and no shared SaaS edition. Those models are not on a roadmap.
    Definition

    What single-tenant means in concrete terms.

    01The instance
    The instance is provisioned for one customer and serves that customer only.
    02What belongs to it
    The application, the data stores behind it and the pipeline tooling belong to that instance.
    03Identity and access
    Credentials, identity configuration and access rules are specific to that instance.
    04Configuration and policy
    Configuration and policy are set for that instance and are not inherited from a shared default that other customers also carry.
    05Upgrades
    Upgrades are applied to that instance on an agreed schedule, rather than to everyone at once.

    Nothing is shared with another customer’s instance. There is no pooled application tier, no pooled database and no shared workload plane between customers.

    Limits

    What is not on offer.

    01No multi-tenant, no shared SaaS
    There is no multi-tenant option and no shared SaaS edition. Those models are not on a roadmap and are not something to ask us for. Single-tenant is the model, wherever the instance runs.
    02The tooling we name
    iTmethods operates GitLab and the surrounding delivery tooling inside the instance. That is the tooling we run, and it is named here because we run it, not as an endorsement of anything else.
    03On security language
    A dedicated instance is a boundary we operate and record. It is not a guarantee about risk, and we will not describe it as one.
    04In development, not available
    Air-gapped deployment and sovereign deployment are both in development. Neither is available today. Neither is being sold. We are not offering a date, because we do not have one worth publishing.
    Boundaries

    What stays inside your boundary, and what does not.

    The shape of the boundary is below. The exact contents are written down for each deployment and agreed before anything runs.

    Inside

    Inside the customer boundary

    • Source code, repositories and branches held in the instance.
    • Build artifacts and pipeline output produced in the instance.
    • The evidence and records generated by work in the instance.
    • Identity and access configuration for the instance.
    • Configuration and policy set for the instance.
    Outside

    Outside the customer boundary

    • Where the instance runs in infrastructure iTmethods operates, that infrastructure is ours. Where it runs in the customer’s own cloud account, the account is theirs and the billing relationship with the cloud provider is theirs. Which of the two applies is written into the deployment record.
    • Operational telemetry required to keep the instance running is handled by our operations team. What that telemetry contains is listed for each deployment.
    • Named operations staff hold administrative access in order to run the platform. Who they are is recorded, and access is reviewed.
    • Platform releases originate with us, not with the customer, and are applied under the change process below.

    If an item matters to a regulator or an internal control, it is written into the deployment record for that instance rather than left to a general statement on a website.

    Change control

    How a change to the deployment is proposed, agreed and recorded.

    Four steps. No verbal changes, and no changes applied because they seemed obvious at the time.

    01

    Proposed

    A change is raised in writing as a merge request in the GitLab instance we operate. It states what is changing, why, what it touches, and how it would be reversed.

    02

    Reviewed

    Named owners on both sides review it. The customer has a named owner and so do we. A change that alters the boundary, access or evidence handling is flagged as such at this point rather than later.

    03

    Agreed

    The change is applied after both named owners approve it. Emergency work follows the same path, with the approval recorded as soon as the situation allows and never quietly skipped.

    04

    Recorded

    The merge request is the record. It carries the date, the approvers, the reasoning and the state before and after. A reviewer can read the sequence of changes to an instance without asking us to reconstruct it.

    Deployment

    Ask about the model, not the roadmap.

    We can walk you through the single-tenant instance in detail, including the parts that sit outside your boundary. If your requirement is air-gapped or sovereign, say so early and we will tell you plainly where that work stands.