Job Retry And Failure Rules

InfraLynx job execution must fail predictably and avoid uncontrolled retry loops.

Retry rules

• retry limits are explicit and job-scoped
• workers requeue failed jobs only while retry allowance remains
• retry count is recorded on the job record
• repeated failure does not create a duplicate job record

Failure rules

• terminal failure moves the job to failed
• failure details are stored in job result metadata
• job logs must include the failure reason
• audit summaries must record the final failure state

Operational rules

• handlers must be designed for idempotent re-execution
• failure alerting is a follow-up concern and remains outside the baseline runtime
• queue backends must preserve job status and logs across worker restarts