What Checking Every Scheduled Run, Not Just the Latest, Actually Caught
42 scheduled runs so far, 5 of them failed outright, and not one of those 5 ever surfaced on its own. No email, no notification, no red mark anywhere I'd naturally look. Windows Task Scheduler ran the job, the job died, and the task simply... sat there, marked complete, looking exactly like every other successful run unless you went and asked it directly. Incident How it was found How long it sat…
42 scheduled runs were conducted, but unfortunately, 5 of them resulted in outright failures. None of these failed runs were flagged or alerted on their own. Windows Task Scheduler executed the job, but the job failed, and the task remained marked as complete, just like any other successful run unless examined directly. The incident was discovered through a routine check, which was performed 10 hours and 36 hours after the corresponding run.
During these checks, it was noted that one run encountered a DNS timeout exit code of 1, while another ran on time but exited with code 3221225786, leading to a forced kill. The habit that ultimately caught both incidents involved including a "블로그 상황 어때" status check, which now involved querying Get-ScheduledTaskInfo for the most recent few task runs instead of merely reviewing the published post list and assuming the absence of a post meant it had never been published.
This extra step was the key factor in detecting both incidents within hours or days instead of them silently becoming "a post that never happened and nobody noticed." The post list alone would not have revealed the issue, as a missing post appears identical whether it is scheduled for a later time or has already failed. While some might argue that a small blog with near-zero traffic doesn't require such operational discipline, the cost of this check is minimal, just a few seconds, compared to the potential cost of not catching a failure, which could result in a gap in a continuous, publicly displayed record.
The failures themselves are not embarrassing, as they involve unglamorous issues like a sleeping Windows machine missing a DNS lookup after waking up. The real embarrassment would have been discovering the failure three weeks later via a reader asking if the blog had stopped posting. However, the author still has some improvements in mind that haven't been implemented yet.
One significant gap remains that this manual check does not address fully: an unattended machine that is offline for a week could still silently lose posts between checks.
Rewritten by urgent.news; may contain errors — check the original.