rexone-core/config/queue.yml at dev · rex-9/rexone-core · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
# config/queue.yml
# Solid Queue Multi-Workload Concurrency Architecture:
# - Fiber-based worker: Non-blocking cooperative async concurrency for I/O-bound jobs
#   (AI streaming/LLM requests, Notification push/email/ActionCable, Stripe webhook processing, default jobs).
#   Dynamically balances across all queues based on priority: when one area surges, fibers
#   scale to handle it; when quiet, CPU and DB connections stay minimal.
# - Thread-based worker: Isolated preemptive OS threads for sequential recurring maintenance
#   and system reconciliation tasks.
# - (Dedicated Media Container via config/queue.media.yml isolates CPU-bound libvips & FFmpeg).

default: &default
  dispatchers: # move scheduled queue to ready queue when time's up [not immediate jobs]
    - polling_interval: 1 # check scheduled queue every 1s & move to ready when scheduled time's up
      batch_size: 500 # batch move to ready queue
      concurrency_maintenance_interval: 60 # check blocked jobs every min

development:
  <<: *default
  workers:
    # High-concurrency non-blocking Fiber reactor for I/O-bound workloads
    # Dynamically scales across queues: payments -> notifications -> error_notifications -> ai -> default
    - queues: [payments, notifications, error_notifications, ai, default]
      fibers: <%= AppConfig::SOLID_QUEUE_FIBERS %>
      polling_interval: <%= AppConfig::SOLID_QUEUE_POLLING_INTERVAL %>

    # Isolated OS Thread worker for recurring scheduler & maintenance jobs
    - queues: [solid_queue_recurring, default]
      threads: <%= AppConfig::SOLID_QUEUE_MAINTENANCE_THREADS %>
      polling_interval: 1

test:
  <<: *default
  workers:
    - queues: "*"
      threads: 1
      polling_interval: 0.5

production:
  <<: *default
  workers:
    # High-concurrency non-blocking Fiber reactor for I/O-bound workloads
    # Dynamically scales across queues: payments -> notifications -> error_notifications -> ai -> default
    - queues: [payments, notifications, error_notifications, ai, default]
      fibers: <%= AppConfig::SOLID_QUEUE_FIBERS %>
      polling_interval: <%= AppConfig::SOLID_QUEUE_POLLING_INTERVAL %>

    # Isolated OS Thread worker for recurring scheduler & maintenance jobs
    - queues: [solid_queue_recurring, default]
      threads: <%= AppConfig::SOLID_QUEUE_MAINTENANCE_THREADS %>
      polling_interval: 1