streamtime_flink_project (Resource)

A Flink project: the home of a Flink application’s uploaded versions, which Flink instances and jobs run. Creating one returns a key your CI system uses to upload versions to it.

Example Usage

# A Flink project with its own artifact store. Your CI system uploads versions to it
# with the project's CI key, which you create when you set up CI (the project page, or
# `streamtime flink-project rotate-key etl`). Terraform does not manage the key.
resource "streamtime_flink_project" "etl" {
  identifier        = "etl"
  name              = "ETL jobs"
  runtime           = "java"
  cluster           = streamtime_flink.shared.identifier # a Job Workspace
  flink_unit_budget = 2

  metadata = jsonencode({
    entry_class = "com.acme.etl.Main"
    labels      = { team = "data" }
  })

  store = {
    store_type     = "oras"
    url            = "https://ghcr.io"
    authentication = "token"
    # A read-only, expiring token limited to this project's artifacts.
    token = var.ghcr_read_token
  }
}

variable "ghcr_read_token" {
  type      = string
  sensitive = true
}

Schema

Required

  • cluster (String) Identifier of the Flink job workspace (a streamtime_flink with mode = session) the project’s Flink units are booked on.
  • flink_unit_budget (Number) Flink units the project may use on its job workspace, booked from the workspace’s capacity.
  • identifier (String) Unique name of the project: up to 24 lowercase letters, digits and dashes. Changing it creates a new project.
  • name (String) Display name of the project.

Optional

  • description (String) Free-text description.
  • metadata (String) Project defaults and labels, as a JSON object set with jsonencode(): entry_class (the default main class of a JAR), default_args, labels and documentation_url. Values Streamtime fills in are not shown as changes.
  • runtime (String) Language of the application: java, scala or python. It decides which artifacts the project accepts. Changing it creates a new project.
  • store (Attributes) Your own artifact store for the project’s releases. Removing it disconnects the store; registered releases stay listed but cannot be pulled until a store is connected again. (see below for nested schema)

Read-Only

  • artifact_kind (String) Kind of artifact the project accepts, from its runtime.
  • build_tool (String) Tool your CI pipeline builds the artifact with, chosen from the runtime: Maven for java, sbt for scala, pip for python.
  • id (String) Same as identifier. Use the identifier to import.
  • latest_version (String) Most recently uploaded version, if any.

Nested Schema for store

Required:

  • store_type (String) Kind of store: oras (an OCI registry) or jfrog.
  • url (String) The store’s https address, for example https://ghcr.io.

Optional:

  • authentication (String) How Flink instances sign in to the store: none (the default), basic or token.
  • ca_certificate (String) CA certificate (PEM) to trust when the store does not use a public certificate.
  • password (String, Sensitive) Password, for basic sign-in. Use a read-only, expiring credential limited to this project’s artifacts. Streamtime never returns it.
  • token (String, Sensitive) Access token, for token sign-in. Use a read-only, expiring token limited to this project’s artifacts. Streamtime never returns it.
  • username (String, Sensitive) Username, for basic sign-in. Streamtime never returns it, so a change made outside Terraform is not shown.

Read-Only:

  • has_credential (Boolean) Whether Streamtime holds a credential for the store.

Import

Import is supported using the following syntax:

The terraform import command can be used, for example:

# Import by identifier. The CI key is not imported: Streamtime returns it only at creation.
terraform import streamtime_flink_project.etl etl