Send your file →

Your automations can fail
while reporting success.

The expensive failures don't throw errors. The backup writes a file — empty. The report arrives on time — with last week's numbers. The sync says "done" — having synced nothing. Normal monitoring can't see any of this, because normal monitoring checks whether the job ran, not whether it produced.

2026-08-03 · 16:52 · our own machine

This happened to us the day we built this page. Our own trading digest fired on schedule, wrote its output file, and passed every freshness check we had — while serving Friday's data under Monday's timestamp. An AI summary layer then narrated three-day-old prices in the present tense, confidently. Every monitor said healthy, because every monitor judged the file, not the content. The report below is the catch, from the tool we built the same afternoon.

What we check that others don't

MISSING / EMPTYThe output exists and isn't a 0-byte file — the classic backup that has "succeeded" nightly for months.
STALE-FILEWritten recently enough. This is the one check everyone does — necessary, nowhere near sufficient.
STALE-CONTENTThe newest date inside the file matches when it was written. A fresh timestamp over old data passes every normal monitor. This is the check that caught our own incident.
FROZENThe content actually changes between runs. A report rewritten byte-identical for a week is a report nobody is feeding.

The weekly evidence report

This is a real one — the actual catch from our own stack, generated by the same engine that would watch yours. Paths generalized; findings verbatim.

watchpost v0.1.0 · generated 2026-08-03 17:04 · client: our own trading stack

1 PROBLEM(S) FOUND — 3 watched, 2 OK, 0 warning, 1 failing
FAILDaily pricing digestreports/scan_digest.txt

STALE-CONTENT: file written 2026-08-03 but newest date inside it is 2026-07-31 (1 business day(s) behind). Fresh timestamp, old data -- the failure mtime checks cannot see.

written 2026-08-03T16:52 · newest data inside: 2026-07-31
OKRegime logreports/regime_log.csv

fresh file, current content

written 2026-07-31T16:48 · newest data inside: 2026-07-31
OKShadow ledgerreports/shadow_log.csv

fresh file, current content

written 2026-08-03T16:52 · newest data inside: 2026-08-03

Plans

$99/month — Watch
  • Up to 10 watched outputs (reports, backups, exports, syncs)
  • Weekly evidence report to your inbox
  • Same-day email when something goes quiet
$225/month — Watch + fix
  • Everything in Watch
  • We fix what we catch — up to 2 hours of repair a month included
  • Quarterly deep audit of the whole automation stack

Honest scope: this watches outputs on your systems — files, exports, reports. If your stack is entirely inside SaaS tools with no exportable output, this isn't the right fit yet, and we'll say so on the first call rather than after the first invoice.

Start with the free version of the question

Not ready for a retainer? Fair. The one-shot version of this check is free and open source — flatline finds what's already quietly dead. Or send one suspicious export through the free file check right now.

Ask about monitoring   See everything we've built →