2

There should be succeeded/failed/pending status in Dataproc job, of course I could see that in Jobs section under Dataproc on Cloud Console. But, how to visualize all of this status in cloud monitoring/stackdriver?

Already tried scorecard chart and use metric submitted jobs but it gives me 0.013/s output, is it possible to display count all the submitted jobs? e.g.: 100 (when I submit 100 jobs to Dataproc cluster)

I also saw failed jobs metric when creating dashboard in cloud monitoring, but not succeeded job, is it possible to add this metric? or maybe I'm missing some steps (sorry, I'm relatively new in logging and monitoring).

Already follow this documentation: https://cloud.google.com/dataproc/docs/guides/logging and add this as properties when creating dataproc cluster:

dataproc:dataproc.logging.stackdriver.job.driver.enable=true
dataproc:dataproc.logging.stackdriver.enable=true
dataproc:jobs.file-backed-output.enable=true
dataproc:dataproc.logging.stackdriver.job.yarn.container.enable=true
Chairul
  • 77
  • 1
  • 11

2 Answers2

1

Option 1:

  • Under Navigation menu select Logging
  • Under Logs enter (click the down arrow on the right side of the text box then select “Convert to advanced filter”)
resource.type="cloud_dataproc_cluster"
resource.labels.cluster_name=[cluster_name]
jsonPayload.message:"completed with exit code"
  • Click Submit, adjust the time range (Last hour, Last 6 hours, etc)
  • Click “Create Metric” link
  • Enter Name on the Metric Editor
  • Select Type=Counter then click “Create Metric”

You will now see the User-defined metric, click the triple vertical dots on the right side of the metric you’ve just created then click Create alert from metric

Option 2:

Using Custom Metrics to create your own metric.

Ismail
  • 877
  • 1
  • 2
  • 7
1

You can find metrics, such as failed_count, submitted_count etc, for Dataproc jobs here. You can use those to create dashboards for monitoring.

Gaurangi Saxena
  • 226
  • 4