Artifacts

Introduction

Artifacts represent a packaged Apache Flink program. For example, a Java DataStream program packaged as a JAR or a Python Table API program packaged as Python script.

Artifacts are Project-scoped and zoned.

artifact.yaml

_11

kind: Artifact

_11

apiVersion: core/v1alpha1

_11

zone: shared-aws-eu-west-1

_11

metadata:

_11

name: top-speed-windowing

_11

project: default

_11

spec:

_11

type: Jar

_11

jarOptions:

_11

entryClass: org.apache.flink.streaming.examples.windowing.TopSpeedwindowing

_11

mainArgs: ["--topic", "input-topic"]

There are different types of Artifacts and each type comes with some type specific options. Currently, only Artifacts of type Jar are supported.

artifact.yaml

_11

kind: Artifact

_11

apiVersion: core/v1alpha1

_11

zone: shared-aws-eu-west-1

_11

metadata:

_11

name: top-speed-windowing-v1

_11

project: default

_11

spec:

_11

type: Jar

_11

jarOptions:

_11

entryClass: org.apache.flink.streaming.examples.windowing.TopSpeedwindowing

_11

mainArgs: ["--topic", "input-topic"]

Lifecycle & Status

Artifacts are created in phase WaitingForUpload.

artifact.yaml

_6

kind: Artifact

_6

apiVersion: core/v1alpha1

_6

...

_6

status:

_6

phase: WaitingForUpload

_6

uri: ""

Once a file was uploaded, the status changes to Ready and the state.uri field is set. During uploading the state changes to Uploading. Should the upload fail or is aborted, the status changes back to WaitingForUpload.

artifact.yaml

_6

kind: Artifact

_6

apiVersion: core/v1alpha1

_6

...

_6

status:

_6

phase: Ready

_6

uri: s3://...

Once an Artifact is Ready it is immutable and no further upload is accepted. Jobs that reference this Artifact can be started now.

status.phase

flowchart TD; WaitingForUpload-->Uploading; Uploading-->Ready; Uploading-->WaitingForUpload; Ready-->Deleting; WaitingForUpload-->Deleting; Uploading-->Deleting;