Prompt
What should I use if I need retries and delayed jobs but don't want much ops overhead?
Latest observation
Oct 24, 2026
OpenAI APIWeb search: off
If you want retries + delayed jobs with minimal ops overhead, the best fit is usually a managed queue / background job service rather than running your own worker infrastructure.
Good low-ops options
-
AWS SQS + Lambda / ECS workers
- Built-in retries, visibility timeout, dead-letter queues
- Delay queues and per-message delay
- Very low maintenance if you’re already on AWS
-
Google Cloud Tasks
- Managed task queue with retry/backoff and scheduling
- Great for HTTP-based workers
- Low ops, simple operational model
-
Azure Queue Storage / Service Bus
- Retries, dead-lettering, delayed delivery
- Best if you’re in Azure
-
Managed job platforms
- Examples: QStash, Upstash Redis + workers, Inngest, Trigger.dev
- Often easiest for app-level background jobs and schedules
- Less infrastructure to manage than a full queue + worker stack
If you want the simplest general recommendation
- AWS SQS if you want a standard, durable queue and don’t mind writing workers
- Cloud Tasks if you want easy delayed HTTP jobs with retries
- QStash / Inngest / Trigger.dev if you want the least operational burden overall
What to avoid if you want low ops
- Self-hosted RabbitMQ, Redis queues (like Sidekiq/Celery with your own Redis), or custom cron/worker setups
These work well, but they add more operational responsibility.
If you tell me your cloud provider and whether your jobs are HTTP-based or worker-based, I can suggest the best specific choice.