Background
At the moment, Cloud Pipeline allows viewing the nodes utilization reports in the GUI. But in some cases, users would like to export this information for further usage in an automated manner.

Approach

• Implement a new command in the CLI, which shall export the CSV file with the node usage metrics:
• Command shall have the following signature:

pipe cluster monitor --instance-id <CLOUD_INSTANCE_ID> \
                                   --run-id <RUN_ID> \
                                   --from <DATE_FROM> \
                                   --to <DATE_TO> \
                                   --interval <REPORT_BREAKDOWN_INTERVAL> \
                                   --output <RESULT_FILE_NAME>

• One of --instance-id and --run-id options shall be specified. If a --run-id is set - then the instance id shall be retrieved from the respective job. Both options cannot be set together.
• --from/--to - dates to generate the report:
  • If --to is not set - use current date
  • If --from is not set - use --to minus 1 day
• --interval - same periods list as we have in the GUI:

1m
5m
15m
30m
1h
4h
12h

• --output - where to place the resulting file. If not set - create it in cwd with the auto-generated name using the query parameters: <cluster_monitor_$RUN_ID/$INSTANCE_ID_$FROM_$TO_$INTERVAL.csv>