Quick Start


Install Graphsignal agent by running:

pip install graphsignal

Or clone and install the GitHub repository:

git clone https://github.com/graphsignal/graphsignal.git
python setup.py install


Configure Graphsignal agent by specifying your API key directly or via GRAPHSIGNAL_API_KEY environment variable.

import graphsignal


To get an API key, sign up for a free account at graphsignal.com. The key can then be found in your account's Settings / API Keys page.

To track deployments, versions and environments separately, specify a deployment parameter, e.g. graphsignal.configure(deployment='my-app-prod-v1').


Use the following examples to integrate Graphsignal agent into your machine learning application. See integration documentation and API reference for full reference.

Graphsignal agent is optimized for production. All executions wrapped with start_trace method will be measured, but only a few will be recorded and profiled to ensure low overhead.


To measure and trace executions, wrap the code with start_trace method.

with graphsignal.start_trace(endpoint='my-model-predict'):
    # function call or code segment

Other integrations are available as well. See integration documentation for more information.


Enable/disable various code profilers depending on the code and model runtime by passing profiler argument to start_trace method. By default profiler is True and Python profiler is enabled. Set to False to disable profiling.

with graphsignal.start_trace(endpoint='my-model-predict', profiler='pytorch'):
    # function call or code segment

The following values are currently supported: True (or python), tensorflow, pytorch, jax, onnxruntime. See integration documentation for more information on each profiler.

Exception tracking

When with context manager is used with start_trace method, exceptions are automatically recorded. For other cases, use EndpointTrace.set_exception method.

Data monitoring

To track data metrics and record data profiles, EndpointTrace.set_data method can be used.

with graphsignal.start_trace(endpoint='my-model-predict') as trace:
    trace.set_data('input', input_data)

The following data types are currently supported: list, dict, set, tuple, str, bytes, numpy.ndarray, tensorflow.Tensor, torch.Tensor.

No raw data is recorded by the agent, only statistics such as size, shape or number of missing values.


After everything is setup, log in to Graphsignal to monitor and analyze execution performance and monitor for issues.


Model serving

import graphsignal



def predict(x):
    with graphsignal.start_trace(endpoint='my-model-predict'):
        return model(x)

Batch job

import graphsignal



for x in data:
    with graphsignal.start_trace(endpoint='my-model-predict', tags=dict(job_id='job1')):
        preds = model(x)

More integration examples are available in examples repo.


Although profiling may add some overhead to applications, Graphsignal only profiles certain executions, automatically limiting the overhead.


To enable debug logging, add debug_mode=True to configure(). If the debug log doesn't give you any hints on how to fix a problem, please report it to our support team via your account.

In case of connection issues, please make sure outgoing connections to https://agent-api.graphsignal.com are allowed.

For GPU profiling, if libcupti agent is failing to load, make sure the NVIDIA® CUDA® Profiling Tools Interface (CUPTI) is installed by running:

/sbin/ldconfig -p | grep libcupti