Skip to content

Bedrock Converse API - Consistent Timeout Issue with Larger models like Anthropic Sonnet 3.5 and Meta Llama 405B

0

And yes it couldn't be any sort of rate-limit because it was my first message in last 48hrs.

Region Name - us-west-2

WARNING:bedrock_converse.utils:Retrying core.llama.bedrock_converse.utils.converse_with_retry_async._retry in 4.0 seconds as it raised ReadTimeoutError: Read timeout on endpoint URL: "https://bedrock-runtime.us-west-2.amazonaws.com/model/meta.llama3-1-405b-instruct-v1%3A0/converse".
Retrying bedrock_converse.utils.converse_with_retry_async._retry in 4.0 seconds as it raised ReadTimeoutError: Read timeout on endpoint URL: "https://bedrock-runtime.us-west-2.amazonaws.com/model/meta.llama3-1-405b-instruct-v1%3A0/converse"
asked a year ago1.4K views
1 Answer
0

Try to add this. Note that it is client side timeout and not server-side.

 config = Config(
     connect_timeout=300,
     read_timeout=300,
 )
 bedrock_client = boto3.client(service_name='bedrock-runtime', config=config)
AWS
answered a year ago
EXPERT
reviewed a year ago
  • I tried this and it only extends the amount of time I must wait for the request to be retried. Instead of waiting 60 seconds, this setting makes you wait 5 minutes for a request that normally takes 5 - 15 seconds. (Note that I'm testing from EC2 instances in the same region as the Bedrock endpoint.)

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.