- Newest
- Most votes
- Most comments
If you are using bedrock in us-east-1 region, you can try to switch to us-west-2. The resource there is more sufficient. Or you can try bedrock cross-region inference: https://aws.amazon.com/cn/blogs/machine-learning/getting-started-with-cross-region-inference-in-amazon-bedrock/
I'm on eu-central-1. Also will need to be on EU servers going forward for compliance reasons
Hi, can you double-check how many requests you are sending per minute? Also, check your lambda code.
answered 2 years ago
Hi, I'm not even able to send a single request.
can you please check your service quota for AWS bedrock and see if the number of requests per minute is more than the available quota .
answered 2 years ago
Hi, I checked this. The quota is greater than 0, but I'm not even able to send a single request. (haven't been able to use Bedrock either through the API or the dashboard, despite model access being granted)
It appears you're experiencing an issue that several other users have recently encountered with Amazon Bedrock. This doesn't seem to be related to your actual usage or token limits, but rather an unusual problem affecting some accounts.
The issue you're describing is likely due to an unexpected change in on-demand quotas for certain regions and models. In some cases, these quotas were inadvertently set to 0, causing throttling errors even for minimal usage.
To resolve this, you'll need to open a support case with AWS. Even if you don't have a technical support plan, you can raise this through the general or billing support channel. Describe the issue, mentioning that you're getting throttling errors despite having quotas greater than 0 and minimal usage.
AWS support should be able to investigate and correct the quotas for your account. This process typically takes about 48 hours to resolve.
It's worth noting that this is not a typical occurrence and doesn't reflect the normal operation of Bedrock. Once resolved, you should be able to use Claude 3.5 Haiku as intended for your SaaS product backend.
In the meantime, you might consider temporarily using a different region if possible, as some users have reported that certain regions (like London) were still functioning correctly during this issue.
Sources
Bedrock: "Too many requests, please wait before trying again." | AWS re:Post
answered 2 years ago
Relevant content
asked a year ago
- AWS OFFICIALUpdated a year ago
Hi there,
I'm Dennis, a Developer Advocate at AWS. I'm really sorry about the trouble this caused and want to thank you for highlighting the issue. We're aware of the problem and are actively working on a resolution.
I'd like to help you directly with this matter. For privacy reasons, it would be best if you could reach out to me via my official Amazon email at traubd@amazon.com or a DM on LinkedIn: https://www.linkedin.com/in/dennis-traub/
I can then gather more details and personally look into your case.
Thanks again, your feedback helps us fix things faster and better! Dennis
I just realized I used my Builder ID. Switched to my internal AWS profile here on re:Post now. To double-check, click on my name above this comment to confirm I'm legit :-)