Batch

Overview

CloudMock emulates AWS Batch, supporting compute environments, job queues, job definitions, scheduling policies, job submission/management, and resource tagging. Job state transitions are simulated synchronously in test mode.

Supported Operations

Operation	Status	Notes
CreateComputeEnvironment	Supported	MANAGED and UNMANAGED types; state machine: CREATING → VALID
DescribeComputeEnvironments	Supported	Filter by name or ARN
UpdateComputeEnvironment	Supported	Update state or service role
DeleteComputeEnvironment	Supported	Stops lifecycle machine
CreateJobQueue	Supported	Priority-ordered compute environment list
DescribeJobQueues	Supported	Filter by name or ARN
UpdateJobQueue	Supported	Update state, priority, or CE order
DeleteJobQueue	Supported
RegisterJobDefinition	Supported	Multi-revision support per definition name
DescribeJobDefinitions	Supported	Filter by name or status
DeregisterJobDefinition	Supported	Marks revision as INACTIVE
SubmitJob	Supported	Supports dependsOn chains
DescribeJobs	Supported	Returns full job detail
ListJobs	Supported	Filter by queue and status
CancelJob	Supported	Only cancels non-terminal jobs
TerminateJob	Supported	Force-stops running jobs
CreateSchedulingPolicy	Supported	Fair-share scheduling policies
DescribeSchedulingPolicies	Supported	Filter by ARN
UpdateSchedulingPolicy	Supported	Update fair-share parameters
DeleteSchedulingPolicy	Supported
TagResource	Supported	Tag any Batch resource by ARN
UntagResource	Supported
ListTagsForResource	Supported

Job State Machine

SUBMITTED → PENDING → RUNNABLE → STARTING → RUNNING → SUCCEEDED / FAILED

Transitions are simulated with configurable delays. In test mode (instant lifecycle), all transitions happen synchronously.

Quick Start

Node.js

import { BatchClient, CreateComputeEnvironmentCommand, CreateJobQueueCommand, RegisterJobDefinitionCommand, SubmitJobCommand } from '@aws-sdk/client-batch';

const client = new BatchClient({
  endpoint: 'http://localhost:4566',
  region: 'us-east-1',
  credentials: { accessKeyId: 'test', secretAccessKey: 'test' },
});

await client.send(new CreateComputeEnvironmentCommand({
  computeEnvironmentName: 'my-ce',
  type: 'MANAGED',
  state: 'ENABLED',
  computeResources: { type: 'EC2', maxvCpus: 256, instanceTypes: ['m5.xlarge'] },
}));

await client.send(new CreateJobQueueCommand({
  jobQueueName: 'my-queue',
  state: 'ENABLED',
  priority: 10,
  computeEnvironmentOrder: [{ computeEnvironment: 'my-ce', order: 1 }],
}));

await client.send(new RegisterJobDefinitionCommand({
  jobDefinitionName: 'my-job-def',
  type: 'container',
  containerProperties: { image: 'alpine:latest', vcpus: 1, memory: 512 },
}));

const { jobId } = await client.send(new SubmitJobCommand({
  jobName: 'my-job',
  jobQueue: 'my-queue',
  jobDefinition: 'my-job-def:1',
}));
console.log('Submitted job:', jobId);

Python

import boto3

client = boto3.client('batch',
    endpoint_url='http://localhost:4566',
    region_name='us-east-1',
    aws_access_key_id='test',
    aws_secret_access_key='test')

client.create_compute_environment(
    computeEnvironmentName='my-ce',
    type='MANAGED',
    state='ENABLED',
    computeResources={'type': 'EC2', 'maxvCpus': 256, 'instanceTypes': ['m5.xlarge']})

response = client.submit_job(
    jobName='my-job',
    jobQueue='my-queue',
    jobDefinition='my-job-def:1')
print(response['jobId'])

Configuration

cloudmock.yml

services:
  batch:
    enabled: true

Known Differences from AWS

• Jobs are not actually executed on compute resources
• Job status transitions are simulated (not driven by real compute)
• Compute environments do not provision real EC2 instances
• Fair-share scheduling policies are stored but not enforced during job dispatch