Redshift

A local Amazon Redshift control plane — clusters, snapshots, parameter/subnet groups, snapshot schedules, endpoint access, and logging, all as real persisted state with AWS-faithful responses.

fakecloud implements the Amazon Redshift control plane (redshift, the 2012-12-01 Query/XML API) as real persisted state. Unlike RDS or ElastiCache there is no data plane container to run — Redshift's SQL surface is the separate redshift-data API — so every one of the 141 operations is modelled as genuine CRUD with validation, pagination, and AWS-shaped error codes rather than stubs.

Supported features

  • Clusters — CreateCluster / ModifyCluster / DeleteCluster / DescribeClusters and the lifecycle actions (RebootCluster, PauseCluster, ResumeCluster, ResizeCluster, RotateEncryptionKey). New clusters progress straight to available with a synthetic Endpoint (<id>.<random>.<region>.redshift.amazonaws.com:5439), a leader ClusterNode, a cluster public key, and a MultiAZ status, so the Terraform provider's create waiter completes immediately.
  • Parameter groups — create/modify/reset/describe, with DescribeClusterParameters honouring the Source=user filter so only explicitly-modified parameters are returned (no perpetual Terraform drift on engine defaults).
  • Subnet & security groups — full CRUD, ingress authorize/revoke, and tag round-tripping.
  • Snapshots — manual and cluster snapshots carrying a real SnapshotArn, RestoreFromClusterSnapshot, RestoreTableFromClusterSnapshot, batch operations, cross-region snapshot-copy configuration (EnableSnapshotCopy / ModifySnapshotCopyRetentionPeriod), and snapshot schedules with cluster associations surfaced under AssociatedClusters.
  • Endpoint access — Redshift-managed VPC endpoints that inherit the cluster's security groups (or a default), settle to active, and expose a well-formed interface VPC endpoint with a network interface.
  • HSM objects, event subscriptions, usage limits, snapshot copy grants, scheduled actions, authentication profiles, IAM-role attach, partner registration — all real CRUD.
  • Per-cluster logging — EnableLogging / DisableLogging / DescribeLoggingStatus round-trip the LogExports, destination type, and bucket/prefix.
  • Tagging — CreateTags / DeleteTags / DescribeTags resolve the resource ARN to real per-resource tag lists (a ResourceNotFoundFault is returned for an ARN that does not resolve).

Out of scope

  • SQL / data plane — querying a warehouse is the separate redshift-data API, not part of the Redshift control plane.
  • Data sharing and zero-ETL integrations — the Terraform resources for these first create a Redshift Serverless namespace, a separate AWS service fakecloud does not implement.

Example

import boto3
rs = boto3.client("redshift", endpoint_url="http://localhost:4566")

rs.create_cluster(
    ClusterIdentifier="analytics",
    NodeType="ra3.xlplus",
    MasterUsername="admin",
    MasterUserPassword="Passw0rd123",
    ClusterType="single-node",
)

cluster = rs.describe_clusters(ClusterIdentifier="analytics")["Clusters"][0]
# cluster["ClusterStatus"] -> "available"
# cluster["Endpoint"] -> {"Address": "analytics-....redshift.amazonaws.com", "Port": 5439}

rs.create_cluster_snapshot(SnapshotIdentifier="nightly", ClusterIdentifier="analytics")
snap = rs.describe_cluster_snapshots(ClusterIdentifier="analytics")["Snapshots"][0]
# snap["SnapshotArn"] -> "arn:aws:redshift:us-east-1:123456789012:snapshot:analytics/nightly"