New Relic Telemetry SDK

newrelic-telemetry-sdk-python provides a Python library for sending Span and Metric data into New Relic using the Python urllib3 library.

The current SDK supports sending dimensional metrics to the New Relic Metric API and spans to the New Relic Trace API.

Why is this cool?

Dimensional Metrics and traces in New Relic! No agent required.

The telemetry SDK tries to be helpful, so your job of sending telemetry data to New Relic can be done in the right way, easily. We've covered all of the basics for you so you can focus on writing feature code directly related to your business need or interest.

Why would you want to use the telemetry SDK?

We imagine you (or your customers) are interested in seeing the telemetry data, generated by your tool, framework, or code, in New Relic. You can write an exporter to do so! Read below to get started

Installing newrelic_telemetry_sdk

To start, the newrelic-telemetry-sdk package must be installed. To install through pip:

$ pip install newrelic-telemetry-sdk

If that fails, download the library from its GitHub page and install it using:

$ python setup.py install

Reporting Your First Span

Spans provide an easy way to time components of your code. The example code assumes you've set the following environment variables:

• NEW_RELIC_INSERT_KEY

import os
import time
from newrelic_telemetry_sdk import Span, SpanClient

with Span(name='sleep') as span:
    time.sleep(0.5)

span_client = SpanClient(os.environ['NEW_RELIC_INSERT_KEY'])
response = span_client.send(span)
response.raise_for_status()
print('Span sleep sent successfully!')

Reporting Your First Metric

There are 3 different types of metrics:

• GaugeMetric
• CountMetric
• SummaryMetric

Metric Descriptions

Metric Type	Interval Required	Description	Example
Gauge	No	A single value at a single point in time.	Room Temperature.
Count	Yes	Track the total number of occurrences of an event.	Number of errors that have occurred.
Summary	Yes	Track count, sum, min, and max values over time.	The summarized duration of 100 HTTP requests.

Example

The example code assumes you've set the following environment variables:

• NEW_RELIC_INSERT_KEY

import os
import time
from newrelic_telemetry_sdk import GaugeMetric, CountMetric, SummaryMetric, MetricClient

metric_client = MetricClient(os.environ['NEW_RELIC_INSERT_KEY'])

temperature = GaugeMetric("temperature", 78.6, {"units": "Farenheit"})

# Record that there have been 5 errors in the last 2 seconds
errors = CountMetric(name="errors", value=5, interval_ms=2000)

# Record a summary of 10 response times over the last 2 seconds
summary = SummaryMetric(
    "responses", count=10, min=0.2, max=0.5, sum=4.7, interval_ms=2000
)

response = metric_client.send_batch((temperature, errors, summary))
response.raise_for_status()
print("Sent metrics successfully!")

Limitations

The New Relic Telemetry APIs are rate limited. Please reference the documentation for New Relic Metrics API and New Relic Trace API Requirements and Limits on the specifics of the rate limits.