DEPLOY.PREFLIGHT.STATISTICS_UNREAD — the escalation could not read the object a finding names
- Category: safety
- Severity: none — the finding is
undetermined, and a severity would put a number on a check that did not answer - Level: 0
- Stability: stable
- Suites: deploy
- Applies to: PostgreSQL, MySQL
What it means
A deploy run can raise the severity of a lint finding once it sees the live object: a rewrite on a table with twelve rows and one with two hundred million are not the same risk. That raise needs a reading, and this id is what the run reports when the reading did not happen.
Why it is undetermined and not a pass
Because a check that could not look is not a check that found nothing. The un-escalated verdict lint already produced still stands — it was the honest answer before any reading, and it stays the answer now.
The failure direction here is therefore "no escalation", never "wrong severity". The escalator only ever raises, so a reading it could not make leaves the honest verdict in place rather than replacing it with a guess.
Why a DEPLOY.PREFLIGHT.* id
This is a statement about the instance, like every other preflight check makes. Putting it in the lint namespace would create an id that only exists when a database happens to be reachable — a rule id whose existence depends on the environment is one nobody can baseline or look up reliably.
What to do about it
Usually nothing: it means the audit's role could not read the statistics view, or the object was not there to measure. If you want the escalation, give the reading role access to the statistics the message names. If you do not, the lint verdict is still valid on its own terms.