Specs

all_enqueued(Keyword.t()) :: [Oban.Job.t()]

Retrieve all currently enqueued jobs matching a set of criteria.

This is a wrapper around Oban.Testing.all_enqueued/2, see Oban.Testing for more details.

options

Options

See shared options for additional repo-specific options.

Specs

assert_enqueued(Keyword.t(), timeout() | :none) :: true

Assert that a job with particular criteria is enqueued.

This is a wrapper around Oban.Testing.assert_enqueued/2, see Oban.Testing for more details.

options

Options

See shared options for additional repo-specific options.

Specs

drain_jobs([drain_option()]) :: drain_result()

Synchronously execute jobs in one or all queues, from within the current process.

Jobs that are enqueued by a process when Ecto is in sandbox mode are only visible to that process. Calling drain_jobs/1 allows you to control when the jobs are executed and to wait synchronously for all jobs to complete.

This function provides several distinct advantages over the standard Oban.drain_queue/2:

• It can drain jobs across one or all queues simultaneously
• It can return the drained jobs rather than a count summary
• It optimizes the defaults for testing batches, workflows, etc.
• It always uses the SmartEngine to guarantee that Pro worker features work as expected

options

Options

• :queue - an atom specifying the queue to drain, or :all to drain jobs across all queues at once. Defaults to :all when no queue is provided.

• :with_limit — the maximum number of jobs to fetch for draining at once. The limit only impacts how many jobs are fetched at once, not concurrency. When recursion is enabled this is how many jobs are processed per-iteration, and it defaults to 1. Otherwise, there isn't a limit and all available jobs are fetched.

• :with_recursion — whether to draining jobs recursively, or all in a single pass. Either way, jobs are processed sequentially, one at a time. Recursion is required when jobs insert other jobs (e.g. batches), or depend on the execution of other jobs (e.g. workflows). Defaults to true.

• :with_safety — whether to silently catch errors when draining. When false, raised exceptions or unhandled exits are reraised (unhandled exits are wrapped in Oban.CrashError). Defaults to false.

• :with_scheduled — whether to include scheduled or retryable jobs when draining. In recursive mode, which is the default, this will include snoozed jobs, and may lead to an infinite loop if the job snoozes repeatedly. Defaults to true.

• :with_summary — whether to summarize execution results with a map of counts by state, or return a list of each job that was drained. Defaults to true, which returns a summary map.

See shared options for additional repo-specific options.

examples

Examples

Drain all available jobs across all queues:

assert %{completed: 3} = drain_jobs(queue: :all)

Drain and return all executed jobs without a count summary:

assert [%Oban.Job{}, %Oban.Job{}] = drain_jobs(with_summary: false)

Drain including a job that you expect to raise:

assert_raise RuntimeError, fn -> drain_jobs(queue: :risky) end

Drain without recursion to identify snoozed jobs:

assert %{scheduled: 3} = drain_jobs(with_recursion: false)

Drain without staging scheduled jobs:

assert %{completed: 1, scheduled: 0} = drain_jobs(with_scheduled: false)

Drain within a custom prefix:

assert %{completed: 3} = drain_jobs(queue: :default, prefix: "private")

Drain using a specific repo (necessary when calling this function directly):

assert %{completed: 3} = drain_jobs(queue: :default, repo: MyApp.Repo)

Specs

perform_callback(Oban.Worker.t(), callback(), term(), [perform_option()]) ::
  Oban.Worker.result()

Construct and execute a job with a batch handle_* callback.

This helper verifies that the batch worker exports the requested callback handler, along with the standard assertions made by perform_job/3.

examples

Examples

Execute the handle_attempted callback without any args:

assert :ok = perform_callback(MyBatch, :attempted, %{})

Execute the handle_exhausted callback with args:

assert :ok = perform_callback(MyBatch, :exhausted, %{for_account_id: 123})

Specs

perform_chunk(Oban.Worker.t(), [term()], [perform_option()]) ::
  Oban.Pro.Workers.Chunk.result()

Construct a list of jobs and process them with a Chunk worker.

Like perform_job/3, this helper reduces boilerplate when constructing jobs and checks for common pitfalls. Unlike perform_job/3, this helper calls the chunk's process/1 function directly and it won't trigger telemetry events.

examples

Examples

Successfully process a chunk of jobs:

assert :ok = perform_chunk(MyChunk, [%{id: 1}, %{id: 2}])

Process a chunk of jobs with job options:

assert :ok = perform_chunk(MyChunk, [%{id: 1}, %{id: 2}], attempt: 5, priority: 3)

Specs

perform_job(Oban.Worker.t(), term(), [perform_option()]) :: Oban.Worker.result()

See Oban.Testing.perform_job/3.

Specs

process_job(Oban.Worker.t(), term(), [perform_option()]) :: Oban.Worker.result()

See Oban.Testing.perform_job/3.

Specs

refute_enqueued(Keyword.t(), timeout() | :none) :: false

Refute that a job with particular criteria is enqueued.

This is a wrapper around Oban.Testing.refute_enqueued/2, see Oban.Testing for more details.

options

Options

See shared options for additional repo-specific options.

Specs

run_batch([Oban.Job.changeset()], [drain_option()]) :: drain_result()

Insert and execute a complete batch of jobs, along with callbacks, within the test process.

options

Options

Accepts all options for drain_jobs/1, including the repo-specific options listed in shared-options.

examples

Examples

Run a batch:

 ids
 |> Enum.map(&MyBatch.new(%{id: &1}))
 |> MyBatch.new_batch()
 |> run_batch()

Run a batch with a specific repo (necessary when calling this function directly):

run_batch(my_batch, repo: MyApp.Repo, prefix: "private")

Specs

run_chunk([Oban.Job.changeset()], [drain_option()]) :: drain_result()

Insert and execute a chunked jobs within the test process.

This helper overrides the chunk's timeout to force immediate processing of jobs up to the chunk size.

options

Options

Accepts all options for drain_jobs/1, including the repo-specific options listed in shared-options.

examples

Examples

Run jobs in chunks:

1..50
|> Enum.map(&MyChunk.new(%{id: &1}))
|> run_chunk()

Run chunks with an explicit repo (necessary when calling this function directly):

run_chunk(changesets, repo: MyApp.Repo, prefix: "private")

Run chunks only for a specific queue:

run_chunk(changesets, queue: "default")

Specs

run_jobs([Oban.Job.changeset()], [drain_option()]) :: drain_result()

Insert and execute jobs synchronously, within the test process.

This is the basis of all other run_* helpers.

options

Options

Accepts all options for drain_jobs/1, including the repo-specific options listed in shared-options.

examples

Examples

Run a list of jobs:

ids
|> Enum.map(&MyWorker.new(%{id: &1}))
|> run_jobs()

Run jobs with an explicit repo (necessary when calling this function directly):

run_jobs(changesets, repo: MyApp.Repo)

Specs

run_workflow(Oban.Pro.Workers.Workflow.t(), [
  Oban.Pro.Workers.Workflow.new_option() | drain_option()
]) :: drain_result()

Insert and execute a workflow synchronously, within the test process.

This helper augments the workflow with options optimized for testing, but it will still respect all standard workflow options.

options

Options

Accepts all options for drain_jobs/1, including the repo-specific options listed in shared-options.

examples

Examples

Run a basic workflow:

MyFlow.new_workflow()
|> MyFlow.add(:a, MyFlow.new(%{id: 1}))
|> MyFlow.add(:b, MyFlow.new(%{id: 2}), deps: [:a])
|> MyFlow.add(:c, MyFlow.new(%{id: 3}), deps: [:b])
|> run_workflow()

Run a workflow with an explicit repo (necessary when calling this function directly):

run_workflow(workflow, repo: MyApp.Repo)

Specs

start_supervised_oban!([Oban.option()]) :: Registry.key()

Start a supervised Oban instance under the test supervisor.

All valid Oban options are accepted. The supervised instance is registered with a unique reference, rather than the default Oban. That prevents any conflict with Oban instances started by your Application, or with other tests running asynchronously.

Furthermore, this helper automatically adds sandbox allowances for any plugins or queue producers, allowing tests to run async.

After the test finishes the test process will wait for the Oban instance to shut down cleanly.

running-jobs

Running Jobs

By default, the supervised instance won't process any jobs because the poll_interval is set to :infinity. Set the poll_interval to a low value to process jobs normally, without manual draining.

options

Options

Any option accepted by Oban.start_link/1 is acceptable, including the repo-specific options listed in shared options.

examples

Examples

Start a basic supervised instance without any queues or plugins:

name = start_supervised_oban!(repo: MyApp.Repo)
Oban.insert(name, MyWorker.new())

Start the supervisor with a single queue and polling every 10ms:

start_supervised_oban!(repo: MyApp.Repo, poll_interval: 10, queues: [alpha: 10])