rails-jobs

Name: rails-jobs
Author: marckohlbrugge

688stars54forks

How to install rails-jobs?

marckohlbrugge/37signals-skills/rails-jobs

$npx -y skills add marckohlbrugge/37signals-skills --skill rails-jobs

Installs into the current project.

Summary

Apply best practices for Rails background jobs using simple orchestration, idempotency, and safe retries. Use when creating, refactoring, or debugging Active Job and queue-backed workflows.

Files · 1

View on GitHub

SKILL.md

1	# Rails Jobs
2
3	Use for background job design and review. Patterns from Campfire (Resque, 4 thin jobs) and Fizzy (Solid Queue, multi-queue).
4
5	## Defaults
6
7	- Keep jobs shallow; the job body is one line calling a model method: `def perform(card) = card.notify_recipients`.
8	- Make jobs idempotent and safe to retry (e.g. rescue `RecordNotUnique` in the domain op).
9	- Fail loudly on real errors; avoid silent rescue patterns.
10	- Set `ActiveJob::Base.enqueue_after_transaction_commit = true` in app defaults — fixes job-before-data races at the root.
11	- Prefer Solid Queue (database-backed) over Redis-backed queues for new apps; co-locate with Puma (`SOLID_QUEUE_IN_PUMA`) in small deployments.
12
13	## Naming Convention
14
15	- `_later` enqueues, plain name does the work, `_now` when you need an explicit synchronous twin:
16
17	```ruby
18	def notify_recipients # the work — public API the job calls
19	def notify_recipients_later # enqueues NotifyRecipientsJob (often private, called from callbacks)
20	```
21
22	- Capture enqueue-time context as keyword args (`Mention::CreateJob.perform_later(self, mentioner: Current.user)`) — don't rely on `Current` at perform time.
23	- In multi-tenant apps, serialize tenant context into every job: prepend an ApplicationJob extension that captures `Current.account` at enqueue (as a GlobalID) and wraps `perform_now` in `Current.with_account`.
24
25	## Good Patterns
26
27	- Small job payloads; pass records (GlobalID) or IDs, never big object graphs.
28	- Queues split by criticality (`default`, `backend`, `webhooks`) with explicit priority order in queue config.
29	- Separate orchestration jobs from heavy processing jobs.
30	- Stagger recurring jobs at odd minutes (12, 27, 50) to avoid synchronized load spikes.
31	- Bulk enqueue: `due.in_batches { \|batch\| ActiveJob.perform_all_later(batch.map { DeliverJob.new(it) }) }`.
32	- Crash-safe long iteration with `ActiveJob::Continuable`: `step :dispatch` + `find_each(start: step.cursor)` + `step.advance!` — essential for fan-out (webhooks, broadcasts, backfills).
33	- Serialize per-owner work with `limits_concurrency to: 1, key: ->(owner) { owner }` (Solid Queue).
34	- Two-tier async for high fan-out HTTP (web push): one job for the domain event, then an in-process thread pool for the HTTP calls. Resolve all AR data before posting to the pool; drop on `RejectedExecutionError` for backpressure.
35	- For `after_destroy_commit` async work, snapshot needed associations in `before_destroy` — the parent rows may be gone when the job runs.
36
37	## Recurring & Maintenance
38
39	- Recurring tasks invoke plain model class methods (`command: "MagicLink.cleanup"`), not dedicated job classes.
40	- Retention sweeps use `delete_all` on scopes (`stale.delete_all`) — skip callbacks on stale rows.
41	- Schedule cleanup for finished queue jobs, expired tokens, old delivery records.
42	- Prefer reset-on-use over cron resets: check-and-reset inside the domain method (`spend` calls `reset_if_due`), not a scheduled job.
43	- Trim recurring schedules in beta/staging environments to essentials.
44
45	## Error Handling Policy
46
47	- Retry transient failures with `retry_on ..., wait: :polynomially_longer` (timeouts, DNS, `Net::SMTPServerBusy`).
48	- Don't retry permanent failures: rescue, classify by error class/message, log at `:info` severity (it's expected — bad address, full mailbox), and move on. Keep job queue resources for work that can succeed.
49	- Distinguish "destination failed" (record outcome, complete the job) from "our code raised" (mark errored, re-raise for retry) — see rails-webhooks.
50	- Package error taxonomies as concerns and include them into framework jobs (`ActionMailer::MailDeliveryJob.include SmtpDeliveryErrorHandling`).
51
52	## Testing
53
54	- Don't unit-test trivial job classes; test async behavior from model tests with `perform_enqueued_jobs(only: Mention::CreateJob) { ... }` and `assert_enqueued_with(job: ...)`.
55	- Assert the observable outcome (mention created, email sent), not job internals.
56
57	## Red Flags
58
59	- Jobs with complex branching and business rules embedded directly.
60	- Non-idempotent side effects without guards.
61	- Retrying permanently invalid inputs.
62	- Enqueueing jobs before required records are committed.
63	- Reading `Current.user`/`Current.account` inside `perform` without serializing it at enqueue.
64	- Per-item `perform_later` in a loop where `perform_all_later` or Continuable iteration fits.
65	- Cron-style reset jobs when reset-on-use logic is simpler and safer.
66	- Background jobs triggering spurious Turbo broadcasts (wrap in `suppressing_turbo_broadcasts`).

People who liked this also viewed

6 picks

Type

find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 20262.6M27k
grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026633k182k
vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026573k29k
improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026522k182k
setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026454k182k
triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026424k182k

rails-jobs

688stars54forks

View on GitHub

How to install rails-jobs?

marckohlbrugge/37signals-skills/rails-jobs

$npx -y skills add marckohlbrugge/37signals-skills --skill rails-jobs

Installs into the current project.

Summary

Apply best practices for Rails background jobs using simple orchestration, idempotency, and safe retries. Use when creating, refactoring, or debugging Active Job and queue-backed workflows.

Files · 1

View on GitHub

SKILL.md

1	# Rails Jobs
2
3	Use for background job design and review. Patterns from Campfire (Resque, 4 thin jobs) and Fizzy (Solid Queue, multi-queue).
4
5	## Defaults
6
7	- Keep jobs shallow; the job body is one line calling a model method: `def perform(card) = card.notify_recipients`.
8	- Make jobs idempotent and safe to retry (e.g. rescue `RecordNotUnique` in the domain op).
9	- Fail loudly on real errors; avoid silent rescue patterns.
10	- Set `ActiveJob::Base.enqueue_after_transaction_commit = true` in app defaults — fixes job-before-data races at the root.
11	- Prefer Solid Queue (database-backed) over Redis-backed queues for new apps; co-locate with Puma (`SOLID_QUEUE_IN_PUMA`) in small deployments.
12
13	## Naming Convention
14
15	- `_later` enqueues, plain name does the work, `_now` when you need an explicit synchronous twin:
16
17	```ruby
18	def notify_recipients # the work — public API the job calls
19	def notify_recipients_later # enqueues NotifyRecipientsJob (often private, called from callbacks)
20	```
21
22	- Capture enqueue-time context as keyword args (`Mention::CreateJob.perform_later(self, mentioner: Current.user)`) — don't rely on `Current` at perform time.
23	- In multi-tenant apps, serialize tenant context into every job: prepend an ApplicationJob extension that captures `Current.account` at enqueue (as a GlobalID) and wraps `perform_now` in `Current.with_account`.
24
25	## Good Patterns
26
27	- Small job payloads; pass records (GlobalID) or IDs, never big object graphs.
28	- Queues split by criticality (`default`, `backend`, `webhooks`) with explicit priority order in queue config.
29	- Separate orchestration jobs from heavy processing jobs.
30	- Stagger recurring jobs at odd minutes (12, 27, 50) to avoid synchronized load spikes.
31	- Bulk enqueue: `due.in_batches { \|batch\| ActiveJob.perform_all_later(batch.map { DeliverJob.new(it) }) }`.
32	- Crash-safe long iteration with `ActiveJob::Continuable`: `step :dispatch` + `find_each(start: step.cursor)` + `step.advance!` — essential for fan-out (webhooks, broadcasts, backfills).
33	- Serialize per-owner work with `limits_concurrency to: 1, key: ->(owner) { owner }` (Solid Queue).
34	- Two-tier async for high fan-out HTTP (web push): one job for the domain event, then an in-process thread pool for the HTTP calls. Resolve all AR data before posting to the pool; drop on `RejectedExecutionError` for backpressure.
35	- For `after_destroy_commit` async work, snapshot needed associations in `before_destroy` — the parent rows may be gone when the job runs.
36
37	## Recurring & Maintenance
38
39	- Recurring tasks invoke plain model class methods (`command: "MagicLink.cleanup"`), not dedicated job classes.
40	- Retention sweeps use `delete_all` on scopes (`stale.delete_all`) — skip callbacks on stale rows.
41	- Schedule cleanup for finished queue jobs, expired tokens, old delivery records.
42	- Prefer reset-on-use over cron resets: check-and-reset inside the domain method (`spend` calls `reset_if_due`), not a scheduled job.
43	- Trim recurring schedules in beta/staging environments to essentials.
44
45	## Error Handling Policy
46
47	- Retry transient failures with `retry_on ..., wait: :polynomially_longer` (timeouts, DNS, `Net::SMTPServerBusy`).
48	- Don't retry permanent failures: rescue, classify by error class/message, log at `:info` severity (it's expected — bad address, full mailbox), and move on. Keep job queue resources for work that can succeed.
49	- Distinguish "destination failed" (record outcome, complete the job) from "our code raised" (mark errored, re-raise for retry) — see rails-webhooks.
50	- Package error taxonomies as concerns and include them into framework jobs (`ActionMailer::MailDeliveryJob.include SmtpDeliveryErrorHandling`).
51
52	## Testing
53
54	- Don't unit-test trivial job classes; test async behavior from model tests with `perform_enqueued_jobs(only: Mention::CreateJob) { ... }` and `assert_enqueued_with(job: ...)`.
55	- Assert the observable outcome (mention created, email sent), not job internals.
56
57	## Red Flags
58
59	- Jobs with complex branching and business rules embedded directly.
60	- Non-idempotent side effects without guards.
61	- Retrying permanently invalid inputs.
62	- Enqueueing jobs before required records are committed.
63	- Reading `Current.user`/`Current.account` inside `perform` without serializing it at enqueue.
64	- Per-item `perform_later` in a loop where `perform_all_later` or Continuable iteration fits.
65	- Cron-style reset jobs when reset-on-use logic is simpler and safer.
66	- Background jobs triggering spurious Turbo broadcasts (wrap in `suppressing_turbo_broadcasts`).

People who liked this also viewed

6 picks

Type