Hi all,
I’ve deployed an n8n (self-hosted) workflow which uses the Claude 3.5 Sonnet model via AWS Bedrock through the LLM node.
Recently, I’ve started observing some erratic behavior while executing the workflow. Specifically, I’m encountering the following error during the LLM node execution:
Bedrock is unable to process your request.
Error Details:
{
"errorMessage": "Bedrock is unable to process your request.",
"errorDetails": {},
"n8nDetails": {
"time": "5/21/2025, 9:48:28 PM",
"n8nVersion": "1.92.2 (Self Hosted)",
"binaryDataMode": "default",
"cause": {
"name": "ServiceUnavailableException",
"$fault": "server",
"$metadata": {}
}
}
}
This issue occurs both when the workflow is triggered via chat triggers and webhooks.
Notably, I’ve observed this more often during evening hours (post 8 PM) and less frequently during the day. Not sure if that’s relevant, but thought I’d mention it in case it helps identify a pattern.
Has anyone else experienced similar issues with Bedrock integration or Claude 3.5 via the LLM node?
Would really appreciate any insights or suggestions on how to resolve or further debug this.
Thanks in advance!