Telemetry integration for event metrics, logging and error reporting.

Initialization Events

Oban emits the following telemetry event when an Oban supervisor is started:

• [:oban, :supervisor, :init] - when the Oban supervisor is started this will execute

The initialization event contains the following measurements:

• :system_time - The system's time when Oban was started

The initialization event contains the following metadata:

• :conf - The configuration used for the Oban supervisor instance
• :pid - The PID of the supervisor instance

Job Events

Oban emits the following telemetry events for each job:

• [:oban, :job, :start] — at the point a job is fetched from the database and will execute
• [:oban, :job, :stop] — after a job succeeds and the success is recorded in the database
• [:oban, :job, :exception] — after a job fails and the failure is recorded in the database

All job events share the same details about the job that was executed. In addition, failed jobs provide the error type, the error itself, and the stacktrace. The following chart shows which metadata you can expect for each event:

Metadata

• :conf — the executing Oban instance's config
• :job — the executing Oban.Job
• :result — the perform/1 return value, always nil for an exception or crash
• :state — one of :success, :failure, :cancelled, :discard or :snoozed

For :exception events the metadata also includes details about what caused the failure.

• :kind — describes how an error occurred. Here are the possible kinds:

  • :error — from an {:error, error} return value. Some Erlang functions may also throw an :error tuple, which will be reported as :error.
  • :exit — from a caught process exit
  • :throw — from a caught value, this doesn't necessarily mean that an error occurred and the error value is unpredictable

• :reason — a raised exception, wrapped crash, or wrapped error that caused the job to fail. Raised exceptions are passes as is, crashes are wrapped in an Oban.CrashError, timeouts in Oban.TimeoutError, and all other errors are normalized into an Oban.PerformError.

• :stacktrace — the Exception.stacktrace/0 for crashes or raised exceptions. Failures from manual error returns won't contain any application code entries and may have an empty stacktrace.

Engine Events

Oban emits telemetry span events for the following Engine operations:

• [:oban, :engine, :init, :start | :stop | :exception]

• [:oban, :engine, :refresh, :start | :stop | :exception]

• [:oban, :engine, :put_meta, :start | :stop | :exception]

• [:oban, :engine, :check_available, :start | :stop | :exception]

Events for bulk operations also include :jobs for the :stop event:

• [:oban, :engine, :cancel_all_jobs, :start | :stop | :exception]

• [:oban, :engine, :delete_all_jobs, :start | :stop | :exception]

• [:oban, :engine, :fetch_jobs, :start | :stop | :exception]

• [:oban, :engine, :insert_all_jobs, :start | :stop | :exception]

• [:oban, :engine, :prune_jobs, :start | :stop | :exception]

• [:oban, :engine, :rescue_jobs, :start | :stop | :exception]

• [:oban, :engine, :retry_all_jobs, :start | :stop | :exception]

• [:oban, :engine, :stage_jobs, :start | :stop | :exception]

Events for job-level Engine operations also include the job, with the exception of :insert_job, :start, because the job isn't available yet.

• [:oban, :engine, :cancel_job, :start | :stop | :exception]

• [:oban, :engine, :complete_job, :start | :stop | :exception]

• [:oban, :engine, :delete_job, :start | :stop | :exception]

• [:oban, :engine, :discard_job, :start | :stop | :exception]

• [:oban, :engine, :error_job, :start | :stop | :exception]

• [:oban, :engine, :insert_job, :start | :stop | :exception]

• [:oban, :engine, :retry_job, :start | :stop | :exception]

• [:oban, :engine, :snooze_job, :start | :stop | :exception]

Metadata

• :conf — the Oban supervisor's config
• :engine — the module of the engine used
• :job - the Oban.Job in question
• :jobs — zero or more maps with the queue, state for each modified job
• :kind, :reason, :stacktrace — see the explanation in job metadata above

Notifier Events

Oban emits a telemetry span event each time the Notifier is triggered:

• [:oban, :notifier, :notify, :start | :stop | :exception]

Metadata

• :conf — the Oban supervisor's config
• :channel — the channel on which the notification was sent
• :payload - the decoded payload that was sent
• :kind, :reason, :stacktrace — see the explanation in job metadata above

Oban also emits an event when the notifier's sonar, responsible for tracking notifier connectivity, switches connectivity status:

• [:oban, :notifier, :switch]

Metadata

• :conf — see the explanation in metadata above
• :status — one of :isolated, :solitary, or :clustered, see Oban.Notifier.status/1 for details

Plugin Events

All the Oban plugins emit telemetry events under the [:oban, :plugin, *] pattern (where * is either :init, :start, :stop, or :exception). You can filter out for plugin events by looking into the metadata of the event and checking the value of :plugin. The :plugin field is the plugin module that emitted the event. For example, to get Oban.Plugins.Cron specific events, you can filter for telemetry events with a metadata key/value of plugin: Oban.Plugins.Cron.

Oban emits the following telemetry event whenever a plugin executes (be sure to check the documentation for each plugin as each plugin can also add additional metadata specific to the plugin):

• [:oban, :plugin, :init] — when the plugin first initializes
• [:oban, :plugin, :start] — when the plugin beings performing its work
• [:oban, :plugin, :stop] — after the plugin completes its work
• [:oban, :plugin, :exception] — when the plugin encounters an error

The following chart shows which metadata you can expect for each event:

Peer Events

Oban emits a telemetry span event each time an Oban Peer election occurs:

• [:oban, :peer, :election, :start | :stop | :exception]

Metadata

• :conf, :kind, :reason, :stacktrace — see the explanation in notifier metadata above
• :leader — whether the peer is the current leader
• :was_leader — whether the peer was the leader before the election occurred
• :peer — the module used for peering

Queue Events

Oban emits an event when a queue shuts down cleanly, e.g. without being brutally killed. Event emission isn't guaranteed because it is emitted as part of a terminate/1 callback.

• [:oban, :queue, :shutdown]

Metadata

• :conf — see the explanation in metadata above
• :orphaned — a list of job id's left in an executing state because they couldn't finish
• :queue — the stringified queue name

Stager Events

Oban emits an event any time the Stager switches between local and global modes:

• [:oban, :stager, :switch]

Metadata

• :conf — see the explanation in metadata above
• :mode — either local for polling mode or global in the more efficient pub-sub mode

Default Logger

A default log handler that emits structured JSON is provided, see attach_default_logger/0 for usage. Otherwise, if you would prefer more control over logging or would like to instrument events you can write your own handler.

Here is an example of the JSON output for the job:stop event:

{
  "args":{"action":"OK","ref":1},
  "attempt":1,
  "duration":4327295,
  "event":"job:stop",
  "id":123,
  "max_attempts":20,
  "meta":{},
  "queue":"alpha",
  "queue_time":3127905,
  "source":"oban",
  "state":"success",
  "tags":[],
  "worker":"Oban.Integration.Worker"
}

All timing measurements are recorded as native time units but logged in microseconds.

Examples

A handler that only logs a few details about failed jobs:

defmodule MicroLogger do
  require Logger

  def handle_event([:oban, :job, :exception], %{duration: duration}, meta, nil) do
    Logger.warning("[#{meta.queue}] #{meta.worker} failed in #{duration}")
  end
end

:telemetry.attach("oban-logger", [:oban, :job, :exception], &MicroLogger.handle_event/4, nil)

Another great use of execution data is error reporting. Here is an example of integrating with Honeybadger, but only reporting jobs that have failed 3 times or more:

defmodule ErrorReporter do
  def handle_event([:oban, :job, :exception], _, %{attempt: attempt} = meta, _) do
    if attempt >= 3 do
      context = Map.take(meta, [:id, :args, :queue, :worker])

      Honeybadger.notify(meta.reason, metadata: context, stacktrace: meta.stacktrace)
    end
  end
end

:telemetry.attach("oban-errors", [:oban, :job, :exception], &ErrorReporter.handle_event/4, [])