Back to landing page
Operations

Onboarding and Operations

From first call to production go-live — typically in 5–7 business days.

gh0stcloud onboarding is engineer-led with one clear goal: at the end of the process, the first tenant workload is running in production with active monitoring and configured alerts — without the customer having to touch a single infrastructure component.

Onboarding timeline

Onboarding in four phases
Typical onboarding in 5–7 business days — from discovery to production go-live.

Phase 1: Discovery (day 1–2)

What happens:

  • Kick-off call to clarify workload type, tier selection (Starter / Growth / Enterprise) and timeline
  • Technical briefing: which services need to run, storage and networking requirements
  • Contract sign-off and package confirmation

Outcome: Onboarding scope is documented in writing, tenant org is created in Keycloak.

Phase 2: Provisioning (day 1–3)

The gh0stservice team provisions all infrastructure:

  1. Keycloak Organization with initial admin accounts and invite links
  2. Isolated tenant namespace in the target environment — automatically created by our provisioning system
  3. Secrets area in OpenBao with initial credentials, exclusively accessible to your namespace
  4. RBAC — ServiceAccounts and bindings following least-privilege principles
  5. Admission policies — namespace labels, resource quotas and NetworkPolicies active automatically

Outcome: Namespace is live, Keycloak login works, observability starts collecting data.

Phase 3: GitOps setup (day 2–5)

Your Git repository is configured as a Flux CD source. From this point, every commit to your repository is the single mechanism to deploy changes into the cluster — no manual kubectl interaction, no direct cluster access.

What happens:

  • Your repository is configured as a Flux source
  • The first reconcile cycle is triggered and validated
  • The deployment pipeline is tested end-to-end
  • The tenant team runs their first deployment cycle

Outcome: First workload runs in the cluster, Flux CD reconciles automatically on every commit.

Phase 4: Go-live (day 5–7)

Pre-production checklist:

  • All workloads reconcile successfully
  • Observability active: metrics, logs and (optionally) traces visible in the portal
  • Billing attribution visible in the FinOps dashboard
  • Alert rules configured for critical services
  • Tenant admin can independently trigger deployments and view logs
  • SLA tier confirmed and active

Outcome: Production is live, gh0stservice monitors actively, SLA applies from this date.

Day-2 operations

Change typeHow it works
New deployment / updateCommit to tenant repository → Flux CD reconciles automatically
New secretsCreate entry in OpenBao → ESO syncs into namespace
Increase capacityTicket to gh0stservice → quota update via provisioning system
Book dedicated nodeContract extension → dedicated bare-metal server assigned exclusively
Platform updatesgh0stservice communicates maintenance windows, zero-downtime for tenant workloads

No ops team needed on the tenant side

For normal operations, tenants need:

  • Access to a Git repository (own or provided by gh0stservice)
  • Basic Kubernetes/Helm knowledge for writing manifests
  • Access to the gh0stportal

No cluster admin access, no infrastructure knowledge, no on-call shifts for platform issues.

Questions or ready to get started?

Talk to us