SchedulerManager

class pyedb.workflows.job_manager.backend.service.SchedulerManager(scheduler_type: pyedb.workflows.job_manager.backend.job_submission.SchedulerType)

Thin async wrapper around cluster scheduler commands.

Provides live introspection of SLURM and LSF clusters including:

• List of partitions / queues with resource information

• Per-partition: total & free cores, total & free memory

• Global job table (running, pending, etc.)

All methods are coroutines so they can be awaited from the REST layer without blocking the event-loop.

Parameters:

scheduler_typeSchedulerType

Type of scheduler (SLURM or LSF only)

Raises:

ValueError

If scheduler_type is not SLURM or LSF

Overview

Methods

get_partitions	Get list of scheduler partitions/queues with resource information.
get_jobs	Get global job table (all users).

Attributes

scheduler_type

Import detail

from pyedb.workflows.job_manager.backend.service import SchedulerManager

Attribute detail

SchedulerManager.scheduler_type

Method detail

async SchedulerManager.get_partitions() → List[Dict[str, Any]]

Get list of scheduler partitions/queues with resource information.

Returns:

List[Dict[str, Any]]

List of partition dictionaries with keys: - name: Partition/queue name - cores_total: Total available cores - cores_used: Currently used cores - memory_total_gb: Total memory in GB - memory_used_gb: Currently used memory in GB

Raises:

RuntimeError

If scheduler command execution fails

async SchedulerManager.get_jobs() → List[Dict[str, Any]]

Get global job table (all users).

Returns:

List[Dict[str, Any]]

List of job dictionaries with keys: - job_id: Scheduler job ID - partition: Partition/queue name - user: Job owner username - state: Job state (RUNNING, PENDING, etc.) - nodes: Number of nodes allocated - cpus: Number of CPUs allocated - memory_gb: Memory allocated in GB

Raises:

RuntimeError

If scheduler command execution fails