Pipelines running but models and table not showing
Closed this issue · 3 comments
Hi,
I have multiple ETL's and while they seem to run correctly, when I go to BQ I see that the tables are not being created nor the models. I have updated made a git pull already but the ETL's while looking correct in CRMint UI and logs are not writing anything on BQ. The ETL's are configured through IBQML and they dont include anythong custom. Have anyone encountered something similar?
I saw an issue sumilar to this one, and I followed the steps described there, and I still have the same issue
Thank you
It's challenging to know what issue is occurring without any sort of error or log.
Is there a way for you to manually run a pipeline on CRMint and then review the logs in the logging component of Google Cloud to evaluate if you are getting an error and what that error is?
Alternatively, I would suggest deploying CRMint in another Google Cloud project and exporting your pipelines to that project.
You can uninstall CRMint from your current project by running crmint cloud uninstall
If I run it manually the first step appears as run correctly, but the create model step fails
When I go to the logs inside the UI I can see why the second step failed
Here are the logs in Cloud for the error
Here are the logs in cloud printing that the creation of the table was successful, previous step to the one above
I installled it using App engine, and those same logs I attached are the same I se eon the UI. If I have to uninstall CRMint and installit again, would I be able to keep the same url to access the UI instead of the default one?
Your BQScriptExecutor is set to dry_run
. You need to disable "Dry Run" for your Product Propensity job.
Here are more details from the Certification Program:
Note: Make sure the dry run setting for any BigQuery script in CRMint is disabled in order to actually execute the query and have it produce output (table, model, etc).
To safeguard you and your customers from incurring any unexpected BigQuery ML charges, any BigQuery scripts run through CRMint can be set to dry run which will estimate the number of bytes read by the query rather than actually running it. Dry runs do not use query slots, and you are not charged for performing a dry run. You can use the estimate returned by a dry run to calculate query costs in the pricing calculator.
To get an estimate of the number of bytes processed and costs associated with the BigQuery ML script, you can enable the dry run setting in CRMint's BigQuery job. To do so, you can click on the job icon or by clicking the 3 vertical dots in the job icon and selecting "edit."
After that, enable the dry run setting by selecting the checkbox and hitting Save.
When a BigQuery dry run enabled script is executed through CRMint, you will get an additional logged item that shows the number of bytes processed. Use this value in conjunction with the pricing calculator to estimate query costs.