A network automation platform runs dozens of concurrent playbook executions, each triggered by a different change request. Operators complain that when they search the aggregated syslog stream for a failed job, they cannot reliably tie individual log lines back to the specific execution that produced them, because messages from parallel runs are interleaved. You are asked to improve the logging strategy so failures can be traced end-to-end. Which change most directly solves this problem?