Hi AWS community..
I'm running into an issue with my service quotas for Anthropic's Claude 3 Opus model on Amazon Bedrock that I wanted to bring attention to.
I'm developing an application that critically relies on utilizing the Claude Opus service due to it being less-prone to error. However, over three weeks ago, my quota limits for on-demand InvokeModel requests and tokens were unexpectedly reduced to zero without any prior notice.
Despite multiple attempts over several weeks to get this resolved through the normal AWS Support channels, I've had no luck so far. The support staff has been polite but ultimately unable to provide a satisfactory resolution or escalate this to someone with the authority to adjust the quotas.
All they could tell me in the end was:
"The team has mentioned that due to high demand, unfortunately we are unable to increase your account limits for Claude 3 Opus model at this time. We expect to revise the customer limits soon. However in the meantime, we encourage you to try some of the other models that bedrock provides. Please refer to the documentation [1]."
This is severely impacting my operations and ability to move forward with product development. As a paying customer, I expected better service, especially considering my anticipated monthly spend on AWS services once fully operational.
I've reopened the case and requested escalations before, but nothing has come of it so far. I keep getting passed around or told I don't qualify since I'm not an Enterprise customer. I have business level support. And I feel like they're just straight up lying to my face. For instance, a colleague of mine who also does not have Enterprise support (and actually has a lower tier of support than I do) was able to get their Claude Opus quotas raised to a reasonable level without issue.
This quota issue is a major roadblock, and I'm feeling extremely frustrated with the lack of urgency and resolution so far, especially seeing others get proper treatment. I wanted to bring more visibility to this situation and get advice from the community on effective escalation paths to finally get this quota limitation addressed by the appropriate decision-makers.
Any help or advice would be appreciated immensely.
Best,
Kenneth W.
Purposeful AI
