Zebrium batch uploads provide a way for grouping one or more related uploads so that they can be monitored and managed later as a unit. Each batch has a unique id used to identify the batch.
Batch uploads are different from service groups:
Service groups provide a semantic connection across the data in uploads when looking for incidents.
Batch uploads manage the overall phases of uploading and processing data in related logs, for example monitoring if a batch is completed, how many lines of data have been ingested so far, the time taken, and so forth.
Batch uploads are integrated into the
ze CLI in the following main ways:
A standalone upload, using the
ze upCLI, automatically has a batch created for it. The batch id is displayed when the upload is finished so progress can be monitored using the
ze batch stateand
ze batch showCLIs, described below.
A set of related uploads, using the
ze upCLI, can be associated with a specifica batch id that has been created earlier using the
ze batch beginCLI. When all the logs for the batch are uploaded, the batch should be completed using
ze batch end, or if there are errors the batch can be cancelled using
ze batch cancel. When
ze batch endis used all the logs for that batch are processed together by Zebrium.
ze batch CLI subcommand allows batch uploads to be created, completed, cancelled and monitored. It has the syntax:
ze batch begin [--url=<url>] [--auth=<auth>] [--batch_id=<batch_id>] ze batch end [--url=<url>] [--auth=<auth>] -batch_id=<batch_id> ze batch cancel [--url=<url>] [--auth=<auth>] -batch_id=<batch_id> ze batch state [--url=<url>] [--auth=<auth>] -batch_id=<batch_id> ze batch show [--url=<url>] [--auth=<auth>] -batch_id=<batch_id>
Upload a log file, on success the new batch id is displayed, usually with a Processing state, meaning the log has been accepted by Zebrium and is being scanned for incidents:
ze up ... --file=myfile.log State for batch upload baxyz1748ca is Processing Sent sucessfully
Monitor the batch until processing completes:
watch ze batch state ... --batch_id=baxyz1748ca
When the batch upload is completed the state will change, typically to Done. For additional information the
ze batch show option can be used:
ze batch show ... --batch_id=baxyz1748ca Batch ID: baxyz1748ca State: Done Created: 2022-06-08T22:58:18Z Completion Time: 2022-06-08T22:59:45Z Expiration Time: 2022-06-10T22:59:45Z Lines: 377943 Bundles Created: 2 Bundles Completed: 2 Upload time: 0:17 min:sec Processing time: 1:20 min:sec
In this output the expiration time refers to the batch upload metadata, not the uploaded logs or any detected incidents.
ze batch begin and
ze batch end subcommands can be used to create a batch upload that spans several linked files. This allows them to be processed as a unit.
Begin a new batch:
ze batch begin ... New batch upload id: baxyz7357473aac1
Upload several logs as part of the same batch, using the
ze up --batch_id=baxyz7357473aac1 ... --file=file1.log ze up --batch_id=baxyz7357473aac1 ... --file=file2.log ze up --batch_id= baxyz7357473aac1 ... --file=file3.log
End the batch:
ze batch end ... --batch_id=baxyz7357473aac1
The batch upload can be monitored as in the previous example, using the
ze batch state and
ze batch show subcommands.
batch_upload.sh script wraps multiple logs with a batch upload in a single step.
bin/batch_upload.sh from the Zebrium GitHub repository here: https://github.com/zebrium/ze-cli
This example uploads all files from
/var/log using a custom host option to
batch_upload.sh -o '--host=myhost' /var/log/*.log
batch_upload.sh -h for full options.
Note: The upload URL and authentication must be supplied on the command line using
-a options if you are not using a
.zerc file. Do not use the
ze --url or
ze --auth options.