20 Limits
Oracle AI Data Platform imposes limits on its resources to ensure efficient and error-free use.
All limits are independently applied. The following examples illustrate how limits are applied:
- Example 1:
- You have 1 AI Data Platform with of 10 workspaces and 5 workspaces enabled for private networks, or
- You have 2 AI Data Platforms with 5 workspaces each (totaling 10 workspaces). One AI Data Platform has 3 workspaces enabled for private networks, and the other has 2 workspaces enabled for private networks (totaling 5 workspaces enabled for private networks).
- Example 2:
- You have 1 all-purpose compute clusters with 100 driver and worker nodes, or
- You have 25 all-purpose compute clusters each with 4 nodes (totalling 100 nodes across all clusters).
The following table lists various numerical limits for Oracle AI Data Platform resources in a single tenancy.
| Resource | Metric | Scope | Limit |
|---|---|---|---|
| AI Data Platform | Maximum number of AI Data Platforms | Region | 100 |
| Workspace |
Maximum number of workspaces Includes default workspace and workspaces enabled for private network |
Region | 100 |
| Workspaces enabled for private network | Number of workspaces enabled for private network | Region | 5 |
| Catalog | Maximum number of catalogs that can be created | AI Data Platform | 1000 |
| Schema | Maximum number of schema that can be created | Catalog | 10000 |
| Table | Maximum number of tables that can be created | Schema | 100000 |
| Volume | Maximum number of volumes that can be created | Schema | 10000 |
| Jobs | Maximum number of jobs that can be created per hour | Workspace | 10000 |
| Jobs: Concurrent tasks | Maximum number of tasks that are running simultaneously | Workspace | 1000 |
| Jobs: Saved jobs | Maximum number of jobs that can be saved | Workspace | 12000 |
| Jobs: Tasks | Maximum number of tasks
Excludes tasks inside a pipeline task. |
Job | 100 |
| Jobs: Nesting level | Maximum levels of nesting | Job | 3 |
| Jobs: Schedule | Minimum frequency | Job | 30 minutes |
| Jobs: Log retention period | Maximum number of days JobRun logs will be retained | AI Data Platform | 30 |
| Workspace file | Maximum number of files
Includes notebooks |
Workspace | 10000 |
| Workspace folder | Maximum number of folders | Workspace | 10000 |
| Workspace file: File size | Maximum size allowed for a single file | Workspace | 500 MB |
| Workspace folder: folder depth | Maximum depth of folders | Workspace | 25 |
| Workspace: Object name length | Maximum length of an object name | Workspace | 500 characters |
| Compute: Clusters | Maximum number of clusters | Region | 100 |
| Compute:Driver & Worker node | Maximum total number of driver and worker nodes | Region | 500 |
| Compute:OCPU per node: (driver or worker) | Maximum number of OCPU | Node (Driver or worker) | 64 |
| Compute: Memory per OCPU | Memory per OCPU | OCPU | Up to 16 times OCPU (in GB) |
| Compute: Block storage per OCPU | Block storage per OCPU | OCPU | 100 times OCPU (in GB) |
| Compute: NVIDIA GPU | Maximum number of NVIDIA GPU | Region | 3 |
| Compute: Logs retention | Maximum number of days compute logs will be retained | AI Data Platform | 30 |