Skip to main content

Amazon Bedrock AgentCore Integration

Foxhound integrates with Amazon Bedrock AgentCore via the OpenTelemetry bridge and a convenience helper that wires everything up in one call.

Installation​

pip install foxhound-ai[opentelemetry] boto3

Usage​

Using the convenience helper​

from foxhound.integrations.opentelemetry import configure_adot_for_foxhound
import boto3

# Set up once at startup — all subsequent Bedrock AgentCore spans are captured
processor = configure_adot_for_foxhound(
    agent_id="bedrock-agent",
    foxhound_endpoint="https://your-foxhound-instance.com",
    api_key="fox_...",
)

# Run your Bedrock AgentCore agent normally
client = boto3.client("bedrock-agent-runtime", region_name="us-east-1")
response = client.invoke_agent(
    agentId="ABCDEF1234",
    agentAliasId="TSTALIASID",
    sessionId="my-session",
    inputText="What is the weather in Seattle?",
)

await processor.flush()

The configure_adot_for_foxhound() helper wires up the TracerProvider and sets it as the global OTel provider in one call — no manual provider setup required.

Using environment variables​

You can supply credentials via environment variables and skip the api_key parameter:

export FOXHOUND_API_KEY="fox_..."
export FOXHOUND_ENDPOINT="https://your-foxhound-instance.com"
processor = configure_adot_for_foxhound(agent_id="bedrock-agent")

What gets traced​

  • Bedrock AgentCore invocations
  • Underlying LLM calls (model, prompt, completion)
  • Tool use and knowledge base queries
  • Agent orchestration steps