9 min read

Service Catalog & Provisioning

Provision approved cloud resources without consoles — with governance guardrails and Whale AI sizing built in.

ProvisioningCatalogAWSGovernanceWhale AI
01

What is the Service Catalog?

The Service Catalog gives engineering and operations teams a governed self-service portal for cloud resources. Instead of granting developers direct access to cloud consoles, you define approved resources (compute, storage, databases, networking) with pre-set guardrails, and teams request them through BlueWhale Stack.

This eliminates console sprawl, enforces tagging and sizing policies, and gives your cloud platform team a complete audit trail of every provisioned resource — without being a bottleneck.

Self-service for teams

Developers request resources via a simple form — no console access, no IAM keys, no Terraform knowledge required.

Governance built in

Tagging policies, size limits, approval workflows, and budget guardrails are enforced at the catalog level.

Whale AI sizing

Whale AI Spark recommends the right instance type and size based on the described workload — reducing over-provisioning from day one.

Full audit trail

Every provisioning request, approval, and lifecycle action is logged in the platform audit trail.

02

Available AWS resources

AWS provisioning is generally available (GA). The following resource types can be provisioned directly from the catalog:

  • EC2 — virtual machines (instance type, AMI, VPC, subnet, security groups, key pair)
  • S3 — storage buckets (region, access control, versioning, encryption, lifecycle policy)
  • RDS — managed databases (engine, version, instance class, storage, multi-AZ, parameter group)
  • VPC — virtual private clouds (CIDR, subnets, route tables, internet gateway, NAT gateway)
  • EFS — managed NFS file systems (throughput mode, lifecycle, encryption, access points)
Note: Azure and GCP catalog items are on the roadmap. Contact your account team for the current availability timeline.
03

Request a resource

  1. 1

    Open the catalog

    Navigate to Service Catalog in the left menu. Browse resource types by category or search by name.

  2. 2

    Select a resource type

    Click the resource type you need (e.g., EC2 Instance). The request form opens with a description, available options, and any platform-wide guardrails visible.

  3. 3

    Ask Whale AI for sizing advice

    Describe your workload in the Whale AI sidebar: “A Node.js API with 50 concurrent users and peak 200 rpm.” Whale AI Spark will recommend an instance type and justify the recommendation with observed workload patterns.

  4. 4

    Complete the form

    Fill in required fields: name, environment (dev/staging/prod), region, and resource-type-specific options. Tag fields marked Required must be completed — these enforce your organisation's tagging policy.

  5. 5

    Submit for approval (if required)

    Requests for production resources or above a defined cost threshold go to an approval queue. Approvers are notified by email and can approve or reject from the notification link or the BlueWhale Stack Approvals dashboard.

  6. 6

    Track provisioning

    Once approved (or immediately for auto-approved requests), provisioning begins. The request shows a live status — Queued → Provisioning → Complete. The resource appears in Inventory as soon as it is live.

04

Approval workflows

Approval rules are configured in Settings → Service Catalog → Approval policies. Rules can require approval based on resource type, environment, estimated monthly cost, or a combination.

  • Auto-approved — provisioning starts immediately; no approval step
  • Manager approval — the requestor's manager (as mapped in your IdP) must approve
  • Team approval — any member of a specified team can approve
  • Multi-level — requires approval from two separate approvers (e.g., team lead + FinOps)
Tip: Set a cost-threshold rule: all requests with an estimated monthly cost above $500 require Finance approval. Whale AI provides the cost estimate at request time using current AWS pricing.
05

Governance guardrails

Platform-level guardrails prevent catalog items from being misconfigured at provisioning time, regardless of what the requestor enters.

  • Mandatory tags — any resource missing required tags is rejected before provisioning starts
  • Instance size limits — prevent t3.nano being used in production, or p4d.24xlarge outside ML environments
  • Region restrictions — limit provisioning to regions your organisation has approved
  • Encryption enforcement — block S3 buckets or RDS instances without encryption at rest
  • Budget guardrails — prevent a team from exceeding their monthly cloud budget via the catalog

Have a question or hit a blocker?

Our support team and account engineers are here to help.

Contact support