Pipecat Monitoring & Observability with OpenTelemetry

Pipecat, an open-source framework for voice AI agents, now supports monitoring and observability through OpenTelemetry, enabling developers to export logs, traces, and metrics to SigNoz for real-time visibility into latency, error rates, and usage trends. The integration, detailed in a new guide, requires Python 3.10+, a SigNoz account, and credentials for Deepgram, Cartesia, and OpenAI, and can be set up with automatic instrumentation and minimal code changes.

Overview This guide walks you through setting up monitoring and observability for Pipecat using OpenTelemetry https://opentelemetry.io/ and exporting logs, traces, and metrics to SigNoz. With this integration, you can observe the performance of various models, capture request/response details, and track system-level metrics in SigNoz, giving you real-time visibility into latency, error rates, and usage trends for your Pipecat applications. Instrumenting Pipecat in your AI applications with telemetry ensures full observability across your voice agent workflows, making it easier to debug issues, optimize performance, and understand user interactions. By leveraging SigNoz, you can analyze correlated traces, logs, and metrics in unified dashboards, configure alerts, and gain actionable insights to continuously improve reliability, responsiveness, and user experience. Prerequisites - A SigNoz Cloud account https://signoz.io/teams/ with an active ingestion key or Self Hosted SigNoz instance https://signoz.io/docs/install/self-host/ - Internet access to send telemetry data to SigNoz Cloud - Python 3.10+ with Pipecat installed - For Python: uv installed for managing Python packages Deepgram Account https://console.deepgram.com/signup for STT Cartesia Account https://play.cartesia.ai/text-to-speech for TTS- OpenAI API Key Monitoring Pipecat For detailed information on instrumenting Pipecat applications with OpenTelemetry, see the Pipecat OpenTelemetry documentation https://docs.pipecat.ai/server/utilities/opentelemetry opentelemetry-tracing . Get started with a sample Pipecat starter project by following the Pipecat quickstart docs https://docs.pipecat.ai/getting-started/quickstart No-code auto-instrumentation is recommended for quick setup with minimal code changes. It's ideal when you want to get observability up and running without modifying your application code and are leveraging standard instrumentor libraries. Step 1: Clone the sample voice agent project and setup dependencies git clone https://github.com/pipecat-ai/pipecat-quickstart.git cd agent-starter-python uv sync Step 2: Setup Credentials Copy .env.example to .env and filling in the required keys: DEEPGRAM API KEY OPENAI API KEY CARTESIA API KEY Step 3: Add Automatic Instrumentation uv pip install opentelemetry-distro opentelemetry-exporter-otlp uv run opentelemetry-bootstrap -a requirements | uv pip install --requirement - Step 4: Instrument your Pipecat application task = PipelineTask pipeline, params=PipelineParams enable metrics=True, Required for some service metrics , enable tracing=True, Enable tracing for this task enable turn tracking=True, Enable turn tracking for this task conversation id="customer-123", Optional - will auto-generate if not provided additional span attributes={"session.id": "abc-123"} Optional - additional attributes to attach to the otel span See this example repo https://github.com/pipecat-ai/pipecat-examples/blob/main/open-telemetry/langfuse/bot.py for more details on how to configure instrumentation. Step 5: Your bot.py should look something like this: Copyright c 2024–2025, Daily SPDX-License-Identifier: BSD 2-Clause License """Pipecat Quickstart Example. The example runs a simple voice AI bot that you can connect to using your browser and speak with it. You can also deploy this bot to Pipecat Cloud. Required AI services: - Deepgram Speech-to-Text - OpenAI LLM - Cartesia Text-to-Speech Run the bot using:: uv run bot.py """ import os from dotenv import load dotenv from loguru import logger print "🚀 Starting Pipecat bot..." print "⏳ Loading models and imports 20 seconds, first run only \n" logger.info "Loading Local Smart Turn Analyzer V3..." from pipecat.audio.turn.smart turn.local smart turn v3 import LocalSmartTurnAnalyzerV3 logger.info "✅ Local Smart Turn Analyzer V3 loaded" logger.info "Loading Silero VAD model..." from pipecat.audio.vad.silero import SileroVADAnalyzer logger.info "✅ Silero VAD model loaded" from pipecat.audio.vad.vad analyzer import VADParams from pipecat.frames.frames import LLMRunFrame logger.info "Loading pipeline components..." from pipecat.pipeline.pipeline import Pipeline from pipecat.pipeline.runner import PipelineRunner from pipecat.pipeline.task import PipelineParams, PipelineTask from pipecat.processors.aggregators.llm context import LLMContext from pipecat.processors.aggregators.llm response universal import LLMContextAggregatorPair from pipecat.processors.frameworks.rtvi import RTVIConfig, RTVIObserver, RTVIProcessor from pipecat.runner.types import RunnerArguments from pipecat.runner.utils import create transport from pipecat.services.cartesia.tts import CartesiaTTSService from pipecat.services.deepgram.stt import DeepgramSTTService from pipecat.services.openai.llm import OpenAILLMService from pipecat.transports.base transport import BaseTransport, TransportParams from pipecat.transports.daily.transport import DailyParams logger.info "✅ All components loaded successfully " load dotenv override=True async def run bot transport: BaseTransport, runner args: RunnerArguments : logger.info f"Starting bot" stt = DeepgramSTTService api key=os.getenv "DEEPGRAM API KEY" tts = CartesiaTTSService api key=os.getenv "CARTESIA API KEY" , voice id="71a7ad14-091c-4e8e-a314-022ece01c121", British Reading Lady llm = OpenAILLMService api key=os.getenv "OPENAI API KEY" messages = { "role": "system", "content": "You are a friendly AI assistant. Respond naturally and keep your answers conversational.", }, context = LLMContext messages context aggregator = LLMContextAggregatorPair context rtvi = RTVIProcessor config=RTVIConfig config= pipeline = Pipeline transport.input , Transport user input rtvi, RTVI processor stt, context aggregator.user , User responses llm, LLM tts, TTS transport.output , Transport bot output context aggregator.assistant , Assistant spoken responses Enable tracing in your PipelineTask task = PipelineTask pipeline, params=PipelineParams enable metrics=True, enable usage metrics=True, , enable tracing=True, Enable tracing for this task enable turn tracking=True, observers= RTVIObserver rtvi , @transport.event handler "on client connected" async def on client connected transport, client : logger.info f"Client connected" Kick off the conversation. messages.append {"role": "system", "content": "Say hello and briefly introduce yourself."} await task.queue frames LLMRunFrame @transport.event handler "on client disconnected" async def on client disconnected transport, client : logger.info f"Client disconnected" await task.cancel runner = PipelineRunner handle sigint=runner args.handle sigint await runner.run task async def bot runner args: RunnerArguments : """Main bot entry point for the bot starter.""" transport params = { "daily": lambda: DailyParams audio in enabled=True, audio out enabled=True, vad analyzer=SileroVADAnalyzer params=VADParams stop secs=0.2 , turn analyzer=LocalSmartTurnAnalyzerV3 , , "webrtc": lambda: TransportParams audio in enabled=True, audio out enabled=True, vad analyzer=SileroVADAnalyzer params=VADParams stop secs=0.2 , turn analyzer=LocalSmartTurnAnalyzerV3 , , } transport = await create transport runner args, transport params await run bot transport, runner args if name == " main ": from pipecat.runner.run import main main Step 6: Run your application with auto-instrumentation OTEL RESOURCE ATTRIBUTES="service.name=<service name " \ OTEL EXPORTER OTLP ENDPOINT="https://ingest.<region .signoz.cloud:443" \ OTEL EXPORTER OTLP HEADERS="signoz-ingestion-key=<your-ingestion-key " \ OTEL EXPORTER OTLP PROTOCOL=grpc \ OTEL TRACES EXPORTER=otlp \ OTEL METRICS EXPORTER=otlp \ OTEL LOGS EXPORTER=otlp \ OTEL PYTHON LOG CORRELATION=true \ OTEL PYTHON LOGGING AUTO INSTRUMENTATION ENABLED=true \ <your run command with opentelemetry-instrument is the name of your service <service name <region : Your SigNoz Cloud region https://signoz.io/docs/ingestion/signoz-cloud/overview/ endpoint <your-ingestion-key : Your SigNoz ingestion key https://signoz.io/docs/ingestion/signoz-cloud/keys/ - Replace <your run command with the actual command you would use to run your application. In this case we would use: uv run opentelemetry-instrument python bot.py Open http://localhost:7860 http://localhost:7860 in your browser and click Connect to start talking to your bot. Code-based instrumentation gives you fine-grained control over your telemetry configuration. Use this approach when you need to customize resource attributes, sampling strategies, or integrate with existing observability infrastructure. Step 1: Clone the sample voice agent project and setup dependencies git clone https://github.com/pipecat-ai/pipecat-quickstart.git cd agent-starter-python uv sync Step 2: Setup Credentials Copy .env.example to .env and filling in the required keys: DEEPGRAM API KEY OPENAI API KEY CARTESIA API KEY Step 3: Install additional OpenTelemetry dependencies uv pip install \ opentelemetry-api \ opentelemetry-sdk \ opentelemetry-exporter-otlp \ opentelemetry-instrumentation-httpx \ opentelemetry-instrumentation-system-metrics Step 4: Import the necessary modules in your Python application Traces: python from opentelemetry.exporter.otlp.proto.http.trace exporter import OTLPSpanExporter Logs: python from opentelemetry.sdk. logs import LoggerProvider, LoggingHandler from opentelemetry.sdk. logs.export import BatchLogRecordProcessor from opentelemetry.exporter.otlp.proto.http. log exporter import OTLPLogExporter from opentelemetry. logs import set logger provider import logging Metrics: python from opentelemetry.sdk.metrics import MeterProvider from opentelemetry.exporter.otlp.proto.http.metric exporter import OTLPMetricExporter from opentelemetry.sdk.metrics.export import PeriodicExportingMetricReader from opentelemetry import metrics from opentelemetry.instrumentation.system metrics import SystemMetricsInstrumentor from opentelemetry.instrumentation.httpx import HTTPXClientInstrumentor Step 5: Set up the OpenTelemetry Span Exporter to send traces directly to SigNoz Cloud python from opentelemetry.exporter.otlp.proto.http.trace exporter import OTLPSpanExporter import os span exporter = OTLPSpanExporter endpoint= os.getenv "OTEL EXPORTER TRACES ENDPOINT" , headers={"signoz-ingestion-key": os.getenv "SIGNOZ INGESTION KEY" }, → SigNoz Cloud trace endpoint with appropriate OTEL EXPORTER TRACES ENDPOINT region https://signoz.io/docs/ingestion/signoz-cloud/overview/ endpoint : https://ingest.<region .signoz.cloud:443/v1/traces → Your SigNoz SIGNOZ INGESTION KEY ingestion key https://signoz.io/docs/ingestion/signoz-cloud/keys/ Step 6: Setup Logs python import logging from opentelemetry.sdk.resources import Resource from opentelemetry. logs import set logger provider from opentelemetry.sdk. logs import LoggerProvider, LoggingHandler from opentelemetry.sdk. logs.export import BatchLogRecordProcessor from opentelemetry.exporter.otlp.proto.http. log exporter import OTLPLogExporter import os resource = Resource.create {"service.name": "<service name "} logger provider = LoggerProvider resource=resource set logger provider logger provider otlp log exporter = OTLPLogExporter endpoint= os.getenv "OTEL EXPORTER LOGS ENDPOINT" , headers={"signoz-ingestion-key": os.getenv "SIGNOZ INGESTION KEY" }, logger provider.add log record processor BatchLogRecordProcessor otlp log exporter Attach OTel logging handler to root logger handler = LoggingHandler level=logging.INFO, logger provider=logger provider logging.basicConfig level=logging.INFO, handlers= handler logger = logging.getLogger name is the name of your service <service name → SigNoz Cloud endpoint with appropriate OTEL EXPORTER LOGS ENDPOINT region https://signoz.io/docs/ingestion/signoz-cloud/overview/ endpoint : https://ingest.<region .signoz.cloud:443/v1/logs → Your SigNoz SIGNOZ INGESTION KEY ingestion key https://signoz.io/docs/ingestion/signoz-cloud/keys/ Step 7: Setup Metrics python from opentelemetry.sdk.resources import Resource from opentelemetry.sdk.metrics import MeterProvider from opentelemetry.exporter.otlp.proto.http.metric exporter import OTLPMetricExporter from opentelemetry.sdk.metrics.export import PeriodicExportingMetricReader from opentelemetry import metrics from opentelemetry.instrumentation.system metrics import SystemMetricsInstrumentor import os resource = Resource.create {"service.name": "<service-name "} metric exporter = OTLPMetricExporter endpoint= os.getenv "OTEL EXPORTER METRICS ENDPOINT" , headers={"signoz-ingestion-key": os.getenv "SIGNOZ INGESTION KEY" }, reader = PeriodicExportingMetricReader metric exporter metric provider = MeterProvider metric readers= reader , resource=resource metrics.set meter provider metric provider meter = metrics.get meter name turn on out-of-the-box metrics SystemMetricsInstrumentor .instrument HTTPXClientInstrumentor .instrument is the name of your service <service name → SigNoz Cloud endpoint with appropriate OTEL EXPORTER METRICS ENDPOINT region https://signoz.io/docs/ingestion/signoz-cloud/overview/ endpoint : https://ingest.<region .signoz.cloud:443/v1/metrics → Your SigNoz SIGNOZ INGESTION KEY ingestion key https://signoz.io/docs/ingestion/signoz-cloud/keys/ 📌 Note: SystemMetricsInstrumentor provides system metrics CPU, memory, etc. , and HTTPXClientInstrumentor provides outbound HTTP request metrics such as request duration. If you want to add custom metrics to your Pipecat application, see Python Custom Metrics . Step 8: Instrument your Pipecat application Copyright c 2024–2025, Daily SPDX-License-Identifier: BSD 2-Clause License """Pipecat Quickstart Example. The example runs a simple voice AI bot that you can connect to using your browser and speak with it. You can also deploy this bot to Pipecat Cloud. Required AI services: - Deepgram Speech-to-Text - OpenAI LLM - Cartesia Text-to-Speech Run the bot using:: uv run bot.py """ import os from dotenv import load dotenv from loguru import logger print "🚀 Starting Pipecat bot..." print "⏳ Loading models and imports 20 seconds, first run only \n" logger.info "Loading Local Smart Turn Analyzer V3..." from pipecat.audio.turn.smart turn.local smart turn v3 import LocalSmartTurnAnalyzerV3 logger.info "✅ Local Smart Turn Analyzer V3 loaded" logger.info "Loading Silero VAD model..." from pipecat.audio.vad.silero import SileroVADAnalyzer logger.info "✅ Silero VAD model loaded" from pipecat.audio.vad.vad analyzer import VADParams from pipecat.frames.frames import LLMRunFrame logger.info "Loading pipeline components..." from pipecat.pipeline.pipeline import Pipeline from pipecat.pipeline.runner import PipelineRunner from pipecat.pipeline.task import PipelineParams, PipelineTask from pipecat.processors.aggregators.llm context import LLMContext from pipecat.processors.aggregators.llm response universal import LLMContextAggregatorPair from pipecat.processors.frameworks.rtvi import RTVIConfig, RTVIObserver, RTVIProcessor from pipecat.runner.types import RunnerArguments from pipecat.runner.utils import create transport from pipecat.services.cartesia.tts import CartesiaTTSService from pipecat.services.deepgram.stt import DeepgramSTTService from pipecat.services.openai.llm import OpenAILLMService from pipecat.transports.base transport import BaseTransport, TransportParams from pipecat.transports.daily.transport import DailyParams from pipecat.utils.tracing.setup import setup tracing logger.info "✅ All components loaded successfully " load dotenv override=True async def run bot transport: BaseTransport, runner args: RunnerArguments : logger.info f"Starting bot" stt = DeepgramSTTService api key=os.getenv "DEEPGRAM API KEY" tts = CartesiaTTSService api key=os.getenv "CARTESIA API KEY" , voice id="71a7ad14-091c-4e8e-a314-022ece01c121", British Reading Lady llm = OpenAILLMService api key=os.getenv "OPENAI API KEY" messages = { "role": "system", "content": "You are a friendly AI assistant. Respond naturally and keep your answers conversational.", }, context = LLMContext messages context aggregator = LLMContextAggregatorPair context rtvi = RTVIProcessor config=RTVIConfig config= pipeline = Pipeline transport.input , Transport user input rtvi, RTVI processor stt, context aggregator.user , User responses llm, LLM tts, TTS transport.output , Transport bot output context aggregator.assistant , Assistant spoken responses setup tracing service name="<service-name ", exporter=span exporter, from initialized Span Exporter in Step 5 console export=False, Set to True for debug output Enable tracing in your PipelineTask task = PipelineTask pipeline, params=PipelineParams enable metrics=True, enable usage metrics=True, , enable tracing=True, Enable tracing for this task enable turn tracking=True, observers= RTVIObserver rtvi , @transport.event handler "on client connected" async def on client connected transport, client : logger.info f"Client connected" Kick off the conversation. messages.append {"role": "system", "content": "Say hello and briefly introduce yourself."} await task.queue frames LLMRunFrame @transport.event handler "on client disconnected" async def on client disconnected transport, client : logger.info f"Client disconnected" await task.cancel runner = PipelineRunner handle sigint=runner args.handle sigint await runner.run task async def bot runner args: RunnerArguments : """Main bot entry point for the bot starter.""" transport params = { "daily": lambda: DailyParams audio in enabled=True, audio out enabled=True, vad analyzer=SileroVADAnalyzer params=VADParams stop secs=0.2 , turn analyzer=LocalSmartTurnAnalyzerV3 , , "webrtc": lambda: TransportParams audio in enabled=True, audio out enabled=True, vad analyzer=SileroVADAnalyzer params=VADParams stop secs=0.2 , turn analyzer=LocalSmartTurnAnalyzerV3 , , } transport = await create transport runner args, transport params await run bot transport, runner args if name == " main ": from pipecat.runner.run import main main Step 9: Run your example bot.py uv run bot.py Open http://localhost:7860 http://localhost:7860 in your browser and click Connect to start talking to your bot. View Traces, Logs, and Metrics in SigNoz Your Pipecat voice agent usage should now automatically emit traces, logs, and metrics. You should be able to view traces in Signoz Cloud under the traces tab: When you click on a trace in SigNoz, you'll see a detailed view of the trace, including all associated spans, along with their events and attributes. You should be able to view logs in Signoz Cloud under the logs tab. You can also view logs by clicking on the “Related Logs” button in the trace view to see correlated logs: When you click on any of these logs in SigNoz, you'll see a detailed view of the log, including attributes: You should be able to see Pipecat related metrics in Signoz Cloud under the metrics tab: When you click on any of these metrics in SigNoz, you'll see a detailed view of the metric, including attributes: Troubleshooting Troubleshooting troubleshooting If you don't see your telemetry data: Verify network connectivity - Ensure your application can reach SigNoz Cloud endpoints Check ingestion key - Verify your SigNoz ingestion key is correct Wait for data - OpenTelemetry batches data before sending, so wait 10-30 seconds after making API calls Try a console exporter — Enable a console exporter locally to confirm that your application is generating telemetry data before it's sent to SigNoz Next Steps You can also check out our custom Pipecat dashboard here https://signoz.io/docs/dashboards/dashboard-templates/pipecat-dashboard/ which provides specialized visualizations for monitoring your Pipecat usage in applications. The dashboard includes pre-built charts specifically tailored for LLM usage, along with import instructions to get started quickly. Related integrations Trace the rest of the stack behind your voice agents: LiveKit observability with OpenTelemetry https://signoz.io/docs/livekit-observability/ - trace voice agent sessions, turn latency, and STT and TTS calls Mastra observability with OpenTelemetry https://signoz.io/docs/mastra-observability/ - trace Mastra agents, workflows, and tool calls LlamaIndex observability with OpenTelemetry https://signoz.io/docs/llamaindex-observability/ - trace RAG queries, retrievers, and index operations Pydantic AI observability with OpenTelemetry https://signoz.io/docs/pydantic-ai-observability/ - trace agent runs, output validation, and model calls Monitor the Anthropic API with OpenTelemetry https://signoz.io/docs/anthropic-monitoring/ - trace Claude requests and break down input, output, and cache token usage Browse all LLM observability integrations https://signoz.io/docs/llm-observability/ to instrument the rest of your stack.