request pathWeb
Serves Studio, the public API, and health endpoints. It does not run schema migrations on every start or replace a dedicated background worker.
FURAZHIR / deployment topology
FURAZHIR’s base production model is a separate single-tenant installation. Web, one-shot migrations, and worker roles are separated, PostgreSQL remains the source of truth, and infrastructure keys belong to the customer.
Components ship as one version but run according to purpose. This simplifies migrations, worker scaling, and readiness diagnosis.
request pathServes Studio, the public API, and health endpoints. It does not run schema migrations on every start or replace a dedicated background worker.
migration checkA one-shot role applies immutable SQL migrations before web and worker are updated. Migration state can be checked without applying changes.
async runtimeAcquires leases and processes outbox, retries, and retention. A heartbeat distinguishes a live HTTP process from a functioning background runtime.
FURAZHIR does not hide database, files, or email in its own SaaS. You choose providers and own their region, access, resilience, and contracts.
requiredA separate database per deployment, constrained application and migration roles, private networking, and a verifiable backup policy.
optionalOptional private object storage with lifecycle controls and dedicated credentials; objects are not made public by default.
customer keysTransport TLS verification and provider keys are configured in the customer environment. Outbound effects pass through an observable outbox.
A running container is only the first signal. Production is ready after migrations, readiness, heartbeat, a restore exercise, and a documented rollback.
bootstrapAdmin token, TOTP, and API key are generated; existing configuration is not silently overwritten, and secrets never enter Git or the image.
verifyLiveness covers the process, readiness covers dependencies, and worker heartbeat covers background work; PostgreSQL, S3, and SMTP are probed separately.
recoverBefore traffic, create a backup, exercise restoration, and retain the previous release for controlled rollback.
FURAZHIR does not promise the same SLA across providers. The choice depends on team maturity, data region, network constraints, and willingness to operate PostgreSQL.
| Option | Good fit | Verify |
|---|---|---|
| Cloud VM + Compose | Fast single-node launch | Backups, private network, disk alerts |
| Managed PostgreSQL | Less manual database work | TLS, region, restore, roles, cost |
| Own cluster | A mature platform team exists | Ingress, secrets, jobs, storage, probes |
| On-premise | An internal perimeter is mandatory | Updates, registry, monitoring, DR |
| Timeweb Cloud | Infrastructure is needed in a chosen Russian region | Actual region of every resource and Compose constraints |
WORKING DATA
The working installation uses PostgreSQL in the customer-selected environment. The demo is only for evaluating the interface and never shares customer data.
Every buyer receives a separate installation, database, and keys.
Backup and restoration are exercised before a live launch.
Storage location is verified for every created resource.
Yes. The package uses Docker and Compose for reproducible startup. Production additionally requires secrets, PostgreSQL, backups, TLS, and observability to be configured.
No mandatory external control plane handles the workload. External services appear only when SMTP, S3, a payment provider, or another integration is connected.
No. Location is only one requirement. Legal grounds, documents, notices, protection measures, retention, contracts, and verified operations are also needed.
PLAN THE ENVIRONMENT
The documentation covers components, requirements, and verification steps. The demo lets you inspect the panel before preparing infrastructure.