Platform capability · Deployment on Demand
A full or partial platform from one deployment file, removed and brought back on request
The platform’s provisioning deploys the platform in four layers from one reviewed deployment file: the network, then the data stores, then the platform services. An operator can stop after any layer to get a partial environment, chooses the AWS region and account for each environment, and removes the environment with the same unchanged tooling. Each service image is identified by its content hash, so a later bring-up reuses the released images and does not rebuild them.
Server sizes, the number of message brokers, availability zones, scaling limits and the deployment policy for each service are all in one reviewed deployment file. Two environments differ by their file, with no changes made by hand in the cloud console, and a later bring-up reads the same file.
Layer 0 creates the network and edge. Layer 1 creates the data stores and secrets. Layers 2 and 3 create the platform services. Each layer passes only sanitized outputs to the next. A deployment that stops at layer 1 is a complete network and data layer with no services running. A real apply stops with an error if it is asked to apply below the next unapplied layer, so a partial deployment cannot delete a later one.
Each layer starts with a fresh preview for that layer. The preview lists the resources that would be created or changed, with a fingerprint of that set of changes. The apply runs only when the changes it sees equal the approved set. A delete, a replacement or an unexpected new resource stops the run before anything is changed.
The same unchanged provisioning image performs the teardown. It takes a database snapshot and waits for it to finish before it removes anything. It refuses to touch the retained state store, images, keys and a fixed keep-list, and it accepts resources that are already gone, so a half-finished teardown can be run again. For a shorter break, pause scales services to zero and stops the database, and resume restores what the pause recorded.
Additions that extend the capability described above.
Checks on every request, a separate database, or dedicated infrastructure
Regional placement of customer data, including the EU
Encrypted backups and point-in-time restore
AWS spend history and a month-end forecast
One record of every customer, app, plan and setting, with overrides for each customer
A solution architect will demo the live service against your domain model.