Skip to content

feat(batch-processing-cli): add --max-image-failure-rate flag#2360

Merged
PawelPeczek-Roboflow merged 4 commits into
mainfrom
mjs/configurable-per-shard-image-failure-rate
May 22, 2026
Merged

feat(batch-processing-cli): add --max-image-failure-rate flag#2360
PawelPeczek-Roboflow merged 4 commits into
mainfrom
mjs/configurable-per-shard-image-failure-rate

Conversation

@maxschridde1494
Copy link
Copy Markdown
Contributor

@maxschridde1494 maxschridde1494 commented May 21, 2026

Summary

  • Adds a new --max-image-failure-rate flag to inference rf-cloud batch-processing process-images-with-workflow and restart-job. Value is a float in [0.0, 1.0]; semantic is "fraction of images per shard that may fail before the worker aborts that shard."

Exposes the per-shard image-failure tolerance knob through the CLI.
Adds the flag to `process-images-with-workflow` and `restart-job`,
threads it through `WorkflowProcessingJobV1` to the wire as
`maxImageFailureRate`, and appends it to the restart override
payload parallel to `maxParallelTasks`.

Pure pass-through plumbing; the threshold is enforced inside the
workflows-data-processor worker (shipped separately in the
batch-processing-services repo). The server is the source of truth
for `[0.0, 1.0]` validation — no client-side range check is added,
matching the existing behavior of other numeric override flags in
this command.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@PawelPeczek-Roboflow PawelPeczek-Roboflow merged commit e1a0c4b into main May 22, 2026
15 of 21 checks passed
@PawelPeczek-Roboflow PawelPeczek-Roboflow deleted the mjs/configurable-per-shard-image-failure-rate branch May 22, 2026 15:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants