Skip to content
Leonardo Cristella edited this page Nov 20, 2019 · 9 revisions

MONIT FOR CRAB/ASO

Scripts

Send data to MONIT-ES for CRAB metrics: https://github.com/dmwm/CRABServer/blob/master/src/script/Monitor/GenerateMONIT.py

Send data to MONIT-ES for ASO metrics: https://github.com/dmwm/AsyncStageout/blob/master/bin/aso_metrics_ora.py

Get and aggregate ASO-FTS data from HDFS example: https://github.com/vkuznet/CMSSpark/blob/master/src/python/CMSSpark/aso_stats.py

Dashboards

CRAB grafana dashboard

ASO grafana dashboard

Reporting for global CMS monitoring project

References:

initial discussion meeting and google doc discussed there

issue tracking in JIRA

google sheet with list of high level attributes to be reported for aggregation in InfluxDB

Plan

CRAB will set the following classAds with the indicated values. What we do not set can be set by STOMP feeding script (aka Benjamin's script) or left to the default setting of unknown. The name of the attrivute and in the google sheet is also indicated for reference

| Monitoring Attribute|    classAd         | Possible values          |
| Tool                | CMS_Tool           | User or HammerCloud  [1] |
| Submission Tool     | CMS_SubmissionTool | CRAB                     |
| Type                | CMS_Type           | Analysis or Test [1]     |
| JobType             | CMS_JobType        | User or xyz   [1]        |
| TaskType            | CMS_TaskType       | cmsRun or script or      |
|                     |                    |  HammerCloud             |
| PrimaryInputType    | CMS_InputType      | Data or MC or None [2]   |
| PrimaryInput        |    not valid       |  not set    [3]          |
| SecondaryInput      |    not valid       |  not set    [3]          |

I'd rather add if input data was read locally or via xrootd.

Notes

[1] HammerCloud "mode" will be detected by finding in the submisison configuration an activity flag which starts with HC i.e. general.activity='HCxyz' see MONIT-28 in which case Type will be set to Test, JobType to xyz

[2] will be set according to primary input data tier

[3] no idea what to do, IMO it is totally useless