PanDA Monitoring

PanDA provides advanced Web based monitoring for different groups of PanDA users: scientists, developers, operators, managers, shifters. Panda monitor can also serve a data source for users scripts and custom automatizing. In this section we basic information, needed to monitor jobs and tasks submitted into PanDA.

Task monitoring

Task is the basic entity creating when a new payload comes to PanDA. There are two views in the PanDA monitoring for navigation over tasks: Tasks list and a Task view. The former view displays the selection of tasks:

There are different parameters could be used together with tasks list to provide narrow selection of tasks:

  • days=<number>, hours=<number> defines the left boundary of the time window used for the query. Right boundary is the current time.

  • date_from=<YYYY-MM-DDThh:mm or YYYY-MM-DD>, date_to=(YYYY-MM-DDThh:mm or YYYY-MM-DD) defines exact time range of modification time for tasks selection

  • endtime_from=<YYYY-MM-DDThh:mm or YYYY-MM-DD>, endtime_to=<YYYY-MM-DDThh:mm or YYYY-MM-DD>, endtimerange=<YYYY-MM-DDThh:mm|YYYY-MM-DDThh:mm> defines time boundaries for task end time.

  • earlierthan=<number>, earlierthandays=<number> defines the right boundary in hours or days of the time window of tasks selection relative to the current time.

  • username=<string> selects tasks by user name of a person who submitted them. This parameter supports asterics, e.g. username=James*.

  • tasktype=<prod, anal> filters tasks by the payload origin - production or analysis.

  • limit=<number> limits the size of the data to be retrieved from the PanDA database in order to serve the current query. This parameter could require some big value (e.g. 100000) in order to deliver more data. By default this value is limited to 20000.

  • display_limit=<number> number of tasks with extended information provided to the query results.

  • status=<failed, done, running, > selects tasks which are in one of the status enlisted here.

  • taskname=<string> filters tasks by its name. This parameter supports asterics, e.g. taskname=shared_pipecheck_20210301T161238Z*.

Here are few examples of such queries:

An individual task is accessible by its ID:

Jobs monitoring

Task view provides links to associated jobs in the “Job status summary” table. However jobs could be also accessed independently to the task view. Jobs list query parameters are the similar to ones as for the tasks list:

An individual job is accessible by its ID:

Retrieving job log

PanDA monitoring provides access to logs generated by a payload or/and correspondent Pilot:

../_images/logs_bigpanda.png

Logs become available when a job in the final state.

Information retrieval

PanDA monitoring could be used as a source of information for user’s scripts and applications. To fetch data in JSON format an &jobs flag should be applied to a query, e.g. https://bigpanda.cern.ch/task/24559935/?json .