When Cumulus fails to properly update a granule's status after completion of an IngestAndPublishGranule workflow, not only does its status remain as "running" (possibly "queued"?) instead of being updated to either "failed" or "completed", but a message is written to a dead-letter queue, which is then immediately pulled from the queue and written to S3 at /sqs.
The Cumulus API exposes a means for recovering those messages and retrying them, which can often lead to successfully updating the granule statuses for the granules in the messages.
We want to enhance the Cumulus CLI to be able to use the API to recover those messages, and then update the Cumulus CLI dependency in package.json to the commit that includes the enhancement.
When Cumulus fails to properly update a granule's status after completion of an IngestAndPublishGranule workflow, not only does its status remain as "running" (possibly "queued"?) instead of being updated to either "failed" or "completed", but a message is written to a dead-letter queue, which is then immediately pulled from the queue and written to S3 at/sqs.
The Cumulus API exposes a means for recovering those messages and retrying them, which can often lead to successfully updating the granule statuses for the granules in the messages.
We want to enhance the Cumulus CLI to be able to use the API to recover those messages, and then update the Cumulus CLI dependency in
package.json
to the commit that includes the enhancement.