How do I know livekit has exhausted the max number...
# ask-aip
proud-lawyer-39815
12/28/2025, 1:27 AMHow do I know livekit has exhausted the max number of retries for a particular function tool?
t
tall-airline-16756
12/28/2025, 1:27 AMThanks for reaching out, Mario. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
tall-airline-16756
12/28/2025, 1:27 AMHello! Our team is operating on reduced capacity over the holidays, so our responses may be delayed.
If your issue is critical/urgent, please react to your original message above with the ‼️ emoji and your issue will be escalated with our team.
Happy Holidays!
tall-airline-16756
12/28/2025, 1:27 AMThanks — you're asking how to tell when LiveKit has exhausted max retries for a given function/tool. Here’s how to detect that, depending on the component:
• Webhooks
◦ LiveKit stops retrying when it reaches the configured
MaxRetriesRetryWaitMinRetryWaitMaxMaxAgeMaxDepthhashicorp/go-retryablehttpp
proud-lawyer-39815
12/28/2025, 1:29 AMum... what about retries of function tool in python? how can I programatically capture that?
t
tall-airline-16756
12/28/2025, 1:30 AMThanks — you're asking how to detect, from Python code, when a LiveKit function tool has exhausted its retries. I don't have docs describing a built-in Python hook for function-tool retry exhaustion, so a couple quick clarifying questions so I can give a precise, runnable solution:
1. Which Python integration are you using for the "function tool"?
◦ LiveKit Python SDK, a custom function running on a webhook, AWS Lambda, or another runner?
1. How are retries currently implemented or configured?
◦ Is retrying handled by LiveKit (server-side), by a library like `retryablehttp`/`tenacity`, or by your own code?
With those answers I can:
• Show code to detect final failure (examples for
tenacityp
proud-lawyer-39815
12/28/2025, 1:30 AMI'm talking about agent function tool
t
tall-airline-16756
12/28/2025, 1:30 AMThanks — you're asking about detecting when an agent function tool in Python has exhausted retries. I’m responding to your point that this is an agent function tool.
Short answer: agent function tools do not expose a built-in retry counter or an exhaustion callback you can read programmatically. The Agents framework handles retries internally at the session/job level, but you cannot query a max-retry event for a specific tool.
What you can do: implement retries inside the tool so your code knows when max retries are hit. Example pattern:
Copy code
from livekit.agents import function_tool, RunContext, ToolError
import asyncio
@function_tool()
async def my_tool(self, context: RunContext):
max_retries = 3
for attempt in range(1, max_retries + 1):
try:
# replace with real operation
result = await self._call_external_api()
return result
except Exception as e:
if attempt == max_retries:
# programmatic detection of exhausted retries
await context.session.say("I've tried multiple times but cannot complete the request.")
raise ToolError("Max retries exceeded after 3 attempts") from e
await asyncio.sleep(2 ** (attempt - 1)) # exponential backoff2 Views