A county permit office deploys an Oracle Function that validates uploaded site plans. The handler sometimes spikes above its configured memory and the invoke fails mid-run. Which statement correctly describes how Oracle Functions treats memory?
Select an answer to reveal the explanation.
Short Explanation
Think of function memory like labeled seats on a bus—only certain sizes exist, and if the passenger will not fit, the ride stops. Oracle Functions offers fixed steps from 128 MB up to 3072 MB (default 128). Blow past the chosen step and the platform terminates that invoke.
Full Explanation
Oracle Functions memory is not a free-form knob. Documented configurations are discrete steps—128, 256, 512, 1024, 2048, or 3072 MB—with 128 MB as the default. The county permit handler must be sized to one of those steps that covers peak working set for site-plan validation; Java runtimes commonly need at least 256 MB. When execution exceeds the configured amount, the platform terminates that invoke rather than paging overflow or continuing under soft pressure.
Treating memory as a continuous slider from tens of megabytes to multiple gigabytes misstates the product: you cannot pick arbitrary integer megabyte values, and Monitoring alone is not the consequence of an overrun. Soft-throttle semantics borrowed from OKE request/limit pairs also do not apply; Functions hard-stop when the configured ceiling is crossed. Overflow-to-Object-Storage paging is not a platform feature for function memory—large artifacts belong in Object Storage by design, but that does not salvage an OOM mid-handler.
Exam caveat: sizing memory far above actual use wastes cost and can still fail if the wrong discrete step is chosen; match the official step to measured demand. Operational check: reproduce the failing invoke with metrics and logs for memory, confirm the configured step in the function settings, bump one discrete step if GC or OOM symptoms appear, and retest the same site-plan payload before changing timeout or gateway settings.