Resource budgets for user-supplied runs
Bound compute and memory consumption before executing a large or malformed workload.
Research note · November 14, 2026Research question
How should a test environment handle a run that exceeds its resource budget?
Method
- Define wall-time, CPU, memory and output-size limits for a local fixture. Record the intended stop behavior and distinguish a budget termination from a physical-system failure.
- Test a workload that completes, one that reaches the time limit and one that exhausts the assigned memory budget. Preserve diagnostic state without claiming the stopped run has a complete trajectory.
- Compare results only across compatible budgets. A tighter CPU limit can alter wall-clock deadlines and should not silently change the physical interpretation of simulated time.
Records to retain
- Resource limit configuration
- Termination cause and time
- Partial output status
- Host and runtime versions
Source basis
Docker documents that containers require explicit resource constraints to bound CPU and memory use. These controls are operational limits, not a complete security design for executing untrusted code.
Interpretation & limits
This note is a test protocol, not an approved Imprint sandbox architecture. Isolation, filesystem access and network policy require separate review before accepting untrusted workloads.
Next experiments
- Exercise budget failures on non-production fixtures.
- Define how incomplete runs appear in exports and comparisons.
References
Technical reference
Implementation details supporting this note, drawn directly from the maintained documentation. These are technical references, not additional papers or new experimental results. Expand a topic to inspect its method, fields and limits.