Google Cloud Storage data source
This article provides an introduction on how Upsolver works with Google Cloud Storage along with a guide on how to create a Google Cloud Storage data source in Upsolver.
Upsolver can read events directly from your Google Cloud Storage, assuming events are partitioned by event date (this partitioning defines the folder structure in the cloud storage).
Upsolver auto-detects the file formats date pattern and shows you a preview of the top files that will be ingested with their corresponding dates. Upsolver also auto-detects when to start the ingestion from.
1. From the Data Sources page, click New.
2. Select Google Cloud Storage.
As you fill in the fields, the top matching files with their dates appear in the Preview pane on the right.
- indicates that the file is recognized and will be included in the data source ingestion.
- indicates that the file will not be included.
You can click Refresh to check on their status as you update fields.
4. (Optional) Enter in the path to the data folder in the Google Cloud Storage bucket. If this is not specified, the data is assumed to be in the top-level of the hierarchy.
Upsolver adds an implicit / before and after the folder.
For example, if you have bucket a, with folder b, Upsolver looks for the path:
The full path displayed is read only. You can use it to review the path Upsolver is using to read from S3.
5. Select or enter in the date pattern of the files to be ingested. This is autodetected but can be modified if required.
The following characters are valid:
yyyy MM dd HH mm
Any other letters or characters must be wrapped in single quotes (e.g.
It is also possible to specify the date as follows:
6. (Optional) Select a time to start ingesting files from. This is usually auto detected, but if there is no preview, the date cannot be established and this defaults to today's date. In this case, set the required start date.
7. (Optional) Under File Name Pattern, select the type of pattern to use (All, Starts With, Ends With, Regular Expression, or Glob Expression) and then enter the file name pattern that specifies the file set to ingest.
If all the files in specified folders are relevant, select All.
The pattern given is matched against the file path starting from the full path shown above.
8. Select the content format. This is typically auto-detected, but you can manually select a format.
may appear. This warning can be ignored for POCs (proof of concept).
For production environments, this indicates that if you run your task retroactively (Start Ingestion From), your compute cluster will process a burst of additional tasks, possibly causing delays in outputs and lookup tables running on this cluster.
To prevent this, go to the Clusters page to edit your cluster and set the Additional Processing Units For Replay to a number greater than 0.
11. (Optional) In addition to the streaming data, you may have initial data. In this case, to list the relevant data, enter a file name prefix (e.g. for DMS loads, enter LOAD) under Initial Load Configuration.
12. (Optional) Under Initial Load Configuration, enter a regex pattern to select the required files.
13. Name this data source.
14. Click Continue to preview your data and review the following metrics:
- Sample Size
- Parsed Successfully
- # of Errors
15. (Optional) If there are any errors, click Back to change the settings as required.
16. Click Create.
You can now use your Google Cloud Storage data source.