VFX-E-1150¶
Title¶
ValidationTimeout
Explanation¶
The isolated validation worker exceeded its 10-second wall-clock budget and was killed; the suite's
validity was never determined. validate_suite runs its YAML/schema evaluation inside a disposable
child process rather than in-process, because a tiny, well-formed-looking YAML input can drive a YAML
scanner into an uninterruptible, near-100%-CPU spin that no in-process CancellationToken can recover
from — only OS-level process termination can. This is the one code in the catalogue where
retryable: true applies to something other than a lifecycle condition: a wall-clock kill can be
provoked by transient load on the host rather than by the suite itself, so the identical call
genuinely might succeed next time.
Common causes¶
- Transient host load — CPU contention from other processes — slowing the worker below the 10-second budget.
- A pathological but not deliberately malicious YAML input that is slow to parse (e.g. deeply nested flow-style collections that stay under the hard size/depth caps). A single over-long line is rejected pre-parse as
VFX-D-1107and never reaches the parser, so it is not a cause here. - Slow worker process startup under load (disk contention, cold JIT/AOT startup).
Fixes¶
- Retry the call — this is one of the few codes in this catalogue where a plain retry can genuinely succeed.
- If it recurs consistently for the same suite, check whether that suite triggers slow YAML parsing even though it stays under the hard size/nesting/alias/line-length caps.
- Reduce concurrent load on the host machine while validating large suite sets.