Closed milo-hyben closed 1 month ago
I had discovered issue with the way we calculate cram proportion size during the seqr billing aggregation. Only sucessfully completed jobs are included. If job fails its' cost is being ignored.
I've tried to run seqr.py script to reload some of the old records. It is failing on this line: https://github.com/populationgenomics/cpg-infrastructure/blob/bf3ef24e1d3663442fe237087461b6a42572d48f/cpg_infra/billing_aggregator/aggregate/seqr.py#L623
By investigation the sql records I have noticed that 2022-01-19 we only had failed jobs:
https://batch.hail.populationgenomics.org.au/batches/7284 https://batch.hail.populationgenomics.org.au/batches/7285 https://batch.hail.populationgenomics.org.au/batches/7317 https://batch.hail.populationgenomics.org.au/batches/7321
Their total cost is is around $80 USD as reported by Hail.
This PR is aiming to fix this discrepancy.
Closing this PR without merging as this is not the right solution for the problem. This PR https://github.com/populationgenomics/metamist/pull/872 is solving it.
I had discovered issue with the way we calculate cram proportion size during the seqr billing aggregation. Only sucessfully completed jobs are included. If job fails its' cost is being ignored.
I've tried to run seqr.py script to reload some of the old records. It is failing on this line: https://github.com/populationgenomics/cpg-infrastructure/blob/bf3ef24e1d3663442fe237087461b6a42572d48f/cpg_infra/billing_aggregator/aggregate/seqr.py#L623
By investigation the sql records I have noticed that 2022-01-19 we only had failed jobs:
https://batch.hail.populationgenomics.org.au/batches/7284 https://batch.hail.populationgenomics.org.au/batches/7285 https://batch.hail.populationgenomics.org.au/batches/7317 https://batch.hail.populationgenomics.org.au/batches/7321
Their total cost is is around $80 USD as reported by Hail.
This PR is aiming to fix this discrepancy.