Deployments and changes

We update your environment in a controlled, traceable and reversible way. No change reaches production without going through a defined process.

How a change is applied

It is described in the environment's versioned configuration.
It is reviewed before being applied.
It is promoted in stages: first in test or staging environments and, when appropriate, to production.
The platform converges automatically to the new declared state.
It is recorded: what changed, who changed it and when.

flowchart LR
    C["1 · Described<br/>versioned configuration"] --> R["2 · Reviewed"]
    R --> P["3 · Staged promotion<br/>test → staging → production"]
    P --> K["4 · The platform converges<br/>to the declared state"]
    K --> L["5 · It is recorded<br/>what · who · when"]
    L -. "reversible: roll back to the previous version" .-> C

No noticeable interruptions

Deployments are carried out through progressive replacement: the new version rolls out gradually while the previous one keeps serving requests. The goal is that you don't notice a cutoff during a routine update.

Reversibility

Since every change is a version of the configuration, rolling back means returning to the previous version. If a change causes unexpected behavior, the rollback is quick and clean.

Major changes

For significant changes (for example, a major version upgrade of your application), we plan it with you and coordinate it within an agreed window. See Maintenance.

Traceability for auditing

The complete change history of your environment is available for audit purposes: every configuration change is immutably recorded in version control.