A.1.RI:5.3 - Separate a process run from a reused number

An operator wants to compare resource use in two reports carrying the same process number. The object of interest is one process run. Both reports concern the same machine boot and PID naming scope; their fields are assumed to describe the indicated process consistently.

The first report gives start time 1,000 ticks after boot; the second gives 2,400 ticks. Linux exposes such a start-time field in /proc/pid/stat. The changed start time excludes continuation of the earlier run, despite the repeated number. Keep the two runs’ resource totals separate; subtracting the earlier total from the later one would not measure one run’s interval use.

If the reports contain only the number and executable name, they can fit either continuation or termination followed by reuse. An available lifecycle event may distinguish them. For ongoing supervision, a retained process reference such as a Linux PID file descriptor can make a fresh number-based inference unnecessary.

This case resolves run continuity. A question about a service continuing through replacement processes uses the service’s own continuation criterion.