Hello people!
I love the Estimated cost feature since its a quick way to, not only get an estimated of the cost, but to compare to past version of the same job (a historic view would be nice!).
My question is:
Mostly, the graph looks like this, which is fine:
But this other graph looks weird, don’t you think?
Context:
Both graphs are from the same job but different runs.
|
First one |
Second one |
| Elapsed time |
Succeeded
|
Running
|
| Elapsed time |
3 hr 59 min
|
3 hr 37 min
|
| machine type |
n1-standard8 |
n1-standard8 |
| job type |
batch |
batch |
| Job region |
us-central1 |
us-central1 |
| Dataflow Prime |
Disabled |
Disabled |
| Runner v2 |
Enabled |
Enabled |
| Dataflow Shuffle |
Enabled |
Enabled |
Current vCPUs
|
8
|
8 |
| Total vCPU time |
63.347 vCPU hr
|
27.761 vCPU hr
|
| Total memory time |
237.551 GB hr
|
104.188 GB hr
|
| Current HDD PD |
25 GB
|
25 GB
|
| Total HDD PD time |
197.959 GB hr
|
87.032 GB hr
|
| Total Shuffle data processed |
11.62 GB
|
4.88 GB
|
| Billable Shuffle data processed |
2.9 GB |
1.22 GB |
–
Best regards
David Regalado
Web | Linkedin | Twitter
Good day @davidregalado25 ,
Welcome back to Google Cloud Community!
Unfortunately, you can’t change this in the dataflow UI, its current visualization is set to Stacked Area Chart and you can only edit the time range for its charts and logs by clicking the dropdown (currently set to MAXTIME). However, you can edit the chart in the metrics explorer and save it to a dashboard by clicking the options button (three dots) and clicking (view in Metrics Explorer). You should be able to see different widget types (E.g. Line chart, stacked bar chart, etc.) that you can choose, and you can try editing the MQL query to your need, in this case, estimated cost per run.
You can use this link for more information: https://cloud.google.com/monitoring/mql/query-editor
https://cloud.google.com/monitoring/mql/query-language
Hope this helps!
1 Like
Hi!
Please, take a closer look, there are spaces with no data in that chart.
–
Best regards
David Regalado
Web | Linkedin | Twitter
