Set rate limits on a queue to control how many workflows start in a given period. Rate limits are global across all DBOS processes. This queue starts at most 50
Rate Limit Queue Execution controls the number of workflows that start within a defined time window by setting global rate limits across all DBOS processes. This ensures that tasks such as API calls or database operations do not exceed thresholds that could cause throttling or system overload. For example, you can configure a queue to start no more than 50 workflows every 30 seconds, protecting against spikes in demand and preserving system stability.
This skill is particularly useful when managing workflows that interact with rate-limited services like OpenAI or Anthropic APIs. It can also prevent overload on internal resources by pacing task execution, combining rate limiting with concurrency controls to fine-tune throughput and parallelism.
This skill is designed for performance marketers and growth leads who automate workflows involving third-party APIs with strict rate limits, such as PPC operators managing ad spend data ingestion or SEO strategists coordinating keyword analysis tools. Agency strategists handling multiple client workflows that query external platforms will also benefit by preventing API rejections and maintaining smooth data flow.
Those responsible for scaling data pipelines that interface with rate-sensitive endpoints will find this skill essential to balance aggressive data collection goals with the operational constraints imposed by external services or internal databases.
Practitioners start by defining a named queue for their workflow tasks, such as "llm_tasks" or "api_tasks." Next, they set the rate limit parameters, specifying the maximum number of workflows to start within a given period, for example, 50 workflows per 30 seconds. After that, they can combine rate limiting with concurrency settings to control both the pace and parallelism of workflow execution.
Finally, they monitor workflow throughput and error rates to adjust limits as needed, ensuring compliance with external API quotas or internal resource capacities while optimizing overall task execution speed.
How does rate limiting interact with concurrency? Rate limiting caps the number of workflows started per time period, while concurrency limits how many run simultaneously; both work together to control load.
Can I set different rate limits for different queues? Yes, each named queue can have its own rate limit configuration tailored to specific service constraints.
What happens if workflows exceed the rate limit? Excess workflows are queued and start only when the rate limit window allows, preventing immediate overload or API rejection.
To apply this skill within Metaflow, attach it to an agent task managing API or database workflows that require controlled pacing. Configure the queue with the desired rate limit parameters to ensure task starts do not exceed thresholds. You can expect smoother integration with rate-limited services and more predictable workflow execution timing. For detailed setup guidance, refer to...
For broader context, see our roundup of claude skills for marketing, and read Claude Code workflows for marketing agencies for related setup guidance.