| Limit |
Value |
Set by |
| Minimum Quick Deploy runtime |
5 minutes |
Platform |
| Default extend amount |
60 minutes |
Platform |
| Max GPUs per job |
Group setting |
Tenant admin |
| Max total GPUs across running jobs |
Group setting |
Tenant admin |
| Max GPUs offered in the form |
The lowest of the queue max, per-job limit and total limit |
Derived |
| Services per job |
One or more |
You |
| Limit |
Value |
| Maximum ports per job |
Configured per installation |
| Maximum port number |
65535 |
| Adding ports after creation |
Not possible |
| Duplicate ports |
Rejected |
| Rate |
Set by |
| Token cost |
System settings |
| Job start cost |
System settings |
| Job cost per minute |
System settings |
| Storage cost per TB-hour |
System settings |
| Tenant multiplier and minimum cost |
Tenant |
| Queue GPU/CPU/memory cost factors |
Queue |
The effective rate is base × queue factor × tenant multiplier, floored at the
tenant’s minimum cost.
|
Value |
| Jobs page sizes |
5, 10, 20 |
| Admin table page sizes |
5, 10 |
| Token history range |
90 days maximum |
|
Behaviour |
| Session lifetime |
Expires; you are asked to sign in again |
| Invitation links |
Single use, time-limited, not revivable |
| Password reset links |
Single use, time-limited |
| Node invites |
Validity date and a maximum number of uses |
|
Behaviour |
| Container filesystem |
Deleted when the job ends |
| Mounted bucket |
Survives |
| Duplicate mount targets |
Rejected at submit |
| Read-only buckets |
Mountable, not writable |
