Cloud provider

Cloud-provider-specific limits and constraints when running e6data.

Limits and constraints that vary by cloud provider.

AWS

Limit	Default	Adjustable
Concurrent EC2 instances per cluster	Limited by AWS account quotas	Yes - request quota increase
VPC peering for In-VPC deployments	Up to AWS default limits	Yes - request via support
S3 bucket access	Same-region recommended	Cross-region works but adds latency/cost
Cross-account IAM trust	Supported via role assumption	Standard AWS configuration

Azure

Limit	Default	Adjustable
Subscription quota for VM cores	Per Azure subscription default	Yes - request quota increase in Azure portal
ADLS Gen2 access	Same-region recommended	Cross-region works but adds latency/cost
Cross-tenant access	Supported via Managed Identity federation	Standard Azure configuration
Network Contributor on VNet	Required for In-VPC internal LB	Configured per cluster

Universal limits across clouds

Limit	Value	Notes
Max executors per cluster	10,000	Practical limit; most clusters are far smaller
Concurrent queries per workspace	Workspace-dependent	Contact support to raise
Result set size per query	100 MB via REST API	Use streaming for larger
Catalog name length	64 characters	Reasonable practical limit
Workspace name length	1–18 characters	Hard limit

Region pairing recommendations

For minimal latency and cost, keep these in the same region:

• e6data workspace
• Underlying data (S3, ADLS Gen2)
• Metastore (Hive, Glue, Unity, Polaris)

Cross-region or cross-cloud access works but introduces:

• Egress costs (often significant for large data scans)
• Latency (10–200ms added per query)
• Reliability dependencies on cross-region networking

See also

• Services and dependencies
• Limitations overview