I've been unable to use Amazon Bedrock for about a week. Every model I try (Claude Haiku 4.5, Nova Lite, Nova Micro) returns the same error:
ThrottlingException: Too many tokens per day, please wait before trying again.
I've never successfully made a single Bedrock API call on this account.
When I check Service Quotas, every "tokens per day" quota for every model shows 0 as the applied account-level value and "Not adjustable." The AWS default values are in the billions but don't seem to apply to my account.
This happens in the Bedrock Playground and via the API. I've tried both on-demand and cross-region inference profiles (e.g. us.anthropic.claude-haiku-4-5-20251001-v1:0). Same error on all of them.
my support case for quota increase went unanswered and unassigned for a couple of days.
Has anyone run into this? Is there an account-level flag that needs to be flipped?
My case number is: 177568844200635
I haven't gotten any response from AWS for the past 7 days :(
I tried all other recommendations on here, but nothing. It seems it's a situation where someone from AWS has to handle from their side.