Closed sjdonado closed 5 months ago
If the job fails we lost all the fetched and processed data, to avoid retrying from scratch a retry by checkpoints/stages is proposed.
A new button "Retry" will call the SSE again /search/job/${jobId} The stages and metadata is stored in memory + KV
/search/job/${jobId}
Likely mitigates this issue:
If the job fails we lost all the fetched and processed data, to avoid retrying from scratch a retry by checkpoints/stages is proposed.
A new button "Retry" will call the SSE again
/search/job/${jobId}The stages and metadata is stored in memory + KVLikely mitigates this issue: