Help us improve the AWS re:Post Knowledge Center by sharing your feedback in a brief survey. Your input can influence how we create and update our content to better support your AWS journey.
Questions tagged with Generative AI on AWS
Innovate faster to reinvent customer experiences and applications
Content language: English
Filter questions
Select tags to filter
Sort by
Sort by most recent
Browse through the questions and answers listed below or filter and sort to narrow down your results.
172 results
Conversation 1 Issues: ...
1
answers
0
votes
43
views
asked a month ago
I am using the Nova Sonic v1 model in the ap-northeast-1 region for our project. ***During live conversations, we have observed that the model occasionally misses or drops some words in the transcript...
1
answers
0
votes
61
views
asked a month ago
As I am trying to prompt any word on the bedrock, it is showing an error regarding -
ThrottlingException
Too many tokens per day, please wait before trying again.
Too many tokens per day, please w...
1
answers
0
votes
154
views
asked a month ago
Hi Everyone,
I’m preparing for the AWS Certified Generative AI Developer – Professional (GDP-C01) exam. Has anyone here taken it or started preparing? I’d love to know what kind of topics or practice...
1
answers
2
votes
260
views
asked 2 months ago
I am using **Nova-2 Sonic** in the **ap-northeast-1** region, and I am regularly encountering errors during live conversations. The system frequently returns the error **“System instability detected.”...
1
answers
0
votes
128
views
asked 2 months ago
Here, in the conversation of words, why doesn't the LLM of Novasonic understand?
**logs-1:
Conversation History**
```
Assistant: Hello! This is Sarah calling from ABC Company. Is this a good time to ...
1
answers
0
votes
46
views
asked 2 months ago
I'm using invoke_model in Bedrock with Llama 4 Maverick.
My prompt format looks like this (as per the docs):
<|begin_of_text|>
<|start_header_id|>system<|end_header_id|>
...system prompt...<|eot_id|...
2
answers
0
votes
93
views
asked 3 months ago
I'm setting up Bedrock with the Qwen3 32B model with thinking enabled, but I'm noticing that the response gets truncated fairly consistently. Is there some sort of setting that I'm missing or is this ...
1
answers
0
votes
204
views
asked 3 months ago
We are deploying multiple foundation models through Bedrock. How can we monitor their performance, token usage, and cost impact in production to optimize usage and prevent overspending?
2
answers
0
votes
408
views
asked 3 months ago
Hi,
Recently we deploy some ai agents using stands SDK to shutdown some of ec2 instances after office hours. The ai agent will run daily but on some days we want to set exception not shutting ec2 ins...
1
answers
0
votes
67
views
asked 3 months ago
We experienced intermittent ServiceUnavailableException errors on October 16, 2025 (09:00-15:00 UTC) when calling the ConverseStream operation for Claude Sonnet 4.5.
**Error Message:**
"ServiceUnavai...
1
answers
0
votes
645
views
asked 4 months ago
When using the Bedrock Agent Core AtlassianOAuth2 credential provider, the service generates a dynamic, one-time redirect_uri with a unique suffix for each authorization request. This conflicts with A...
1
answers
0
votes
283
views
asked 4 months ago