JobTemplate

Source: src/AWS/EMRContainers/JobTemplate.ts

An Amazon EMR on EKS job template — a stored set of StartJobRun values (execution role, release label, job driver, configuration) that can be referenced by ID when starting job runs, optionally with ${placeholder} parameters filled in per run.

Job templates are account-level and fully immutable: any change (including tags, which the tagging API does not support post-create for templates) replaces the template. They pair with the AWS.EMRContainers.StartJobRun binding — a Lambda can start a templated Spark job with just the template ID and parameter values.

Creating Job Templates

A Spark Job Template

import * as AWS from "alchemy/AWS";

const template = yield* AWS.EMRContainers.JobTemplate("EtlTemplate", {
  jobTemplateData: {
    executionRoleArn: jobRole.roleArn,
    releaseLabel: "emr-7.5.0-latest",
    jobDriver: {
      sparkSubmitJobDriver: {
        entryPoint: "s3://my-bucket/scripts/etl.py",
        sparkSubmitParameters: "--conf spark.executor.instances=2",
      },
    },
  },
});

Parameterized Template

const template = yield* AWS.EMRContainers.JobTemplate("Parameterized", {
  jobTemplateData: {
    executionRoleArn: jobRole.roleArn,
    releaseLabel: "emr-7.5.0-latest",
    jobDriver: {
      sparkSubmitJobDriver: { entryPoint: "${EntryPoint}" },
    },
    parameterConfiguration: {
      EntryPoint: { type: "STRING" },
    },
  },
});
// StartJobRun with jobTemplateId + jobTemplateParameters: { EntryPoint: "s3://..." }