Class: SagaForge::Dashboard::SuspendedController

Inherits:
BaseController
  • Object
show all
Defined in:
app/controllers/saga_forge/dashboard/suspended_controller.rb

Overview

Instances (across every saga class) with at least one event that ran out of retries and landed in :failed, the saga is suspended pending a resume or compensate decision.

Constant Summary collapse

CAP =

Bound the scan: at scale there can be a large backlog, so we examine at most CAP (most recently updated first) rather than the whole table.

500

Instance Method Summary collapse

Instance Method Details

#index ⇒ Object 



11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
 
# File 'app/controllers/saga_forge/dashboard/suspended_controller.rb', line 11

def index
  @states = SagaForge::State.suspended.order(updated_at: :desc).limit(CAP).to_a
  @capped = @states.size == CAP

  # Single batched query keyed by saga_forge_state_id, grouped in Ruby,
  # instead of one query per row (N+1 at CAP=500 rows). A state can have
  # more than one failed event; earliest (ledger order) wins as "the"
  # failure to show.
  failed = SagaForge::Event.failed
    .where(saga_forge_state_id: @states.map(&:id))
    .select(:saga_forge_state_id, :event_name, :error, :created_at, :id)
    .ledger_order
    .to_a
  @failed_event = failed.group_by(&:saga_forge_state_id).transform_values(&:first)

  # The list is cross-class, but bulk resume is scoped to one class (the
  # BulkRecoveryJob derives its scope from State.for_saga(klass)), so
  # offer one "resume all" button per distinct class present here.
  @saga_classes = @states.map(&:saga_class).uniq.sort
end