Rate limits keep an API reliable for everyone. Each account gets a cap on how many requests it can make in a given window, so one account can't take out the API for everyone else.
Mux, traditionally, has managed rate limits by giving each account a single bucket of requests. That works for a lot of use cases up to now. But as accounts grow and customers have used Mux in different ways, we've seen where that approach starts to fall short. A runaway script in dev, for example, or a batch job summarizing old content can eat through an account’s limit and start refusing requests in production.
Today we're introducing two levers to give you more granular control to prevent that from happening: per-environment limits and token priority.
Dev and production environments carry different risks. In production, every refused request impacts a real user’s experience. In dev, your audience and your tolerance for throttled experiences is much higher. You're poking at things, running a quick test, fixing a bug, or occasionally writing a script that loops more than you meant it to. That's the whole point of a dev environment.
Up to now, both environments drew from the same bucket, putting production at risk. Now every Mux environment gets its own rate limit bucket. Dev, staging, and production are fully isolated. Which means, a mistake in dev stays in dev, and your end-user is none the wiser.
As of today, this is automatically applied to your Mux environments. There’s nothing you need to do to enable this.
In addition to per-environment limits, we’re giving you a way to get even more granular.
Even inside a single environment, not all API calls are equally important. A user uploading a video needs that call to work now. A Mux Robots job summarizing your back catalog can wait for a second or two.
To address this, we have given every environment two buckets: one for high priority tokens and one for low priority tokens. You pick the priority when you create the token. This allows you to assign a low priority token so that it doesn’t compete with the calls critical to your end-users.
Here’s a good rule of thumb:
You can create high and low priority tokens on your organization's settings page. Existing tokens default to high priority and will keep working the way they always have.
We're also adding the standard X-RateLimit-Limit and X-RateLimit-Remaining headers on every API response. You can use these to monitor consumption in real time and slow down your requests before you ever hit a 429 error.
You can start using these new levers and buckets today. Reach out if you have any questions or comments, we'd love to hear these new rate limit levers are fitting into your workflows.
No credit card required to get started.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。