Instrumentation

Manual instrumentation for OpenTelemetry Python

Instrumentation is the act of adding observability code to an app yourself.

If you’re instrumenting an app, you need to use the OpenTelemetry SDK for your language. You’ll then use the SDK to initialize OpenTelemetry and the API to instrument your code. This will emit telemetry from your app, and any library you installed that also comes with instrumentation.

If you’re instrumenting a library, only install the OpenTelemetry API package for your language. Your library will not emit telemetry on its own. It will only emit telemetry when it is part of an app that uses the OpenTelemetry SDK. For more on instrumenting libraries, see Libraries.

For more information about the OpenTelemetry API and SDK, see the specification.

Setup

First, ensure you have the API and SDK packages:

pip install opentelemetry-api
pip install opentelemetry-sdk

Traces

Acquire Tracer

To start tracing, you’ll need to initialize a TracerProvider and optionally set it as the global default.

from opentelemetry import trace
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import (
    BatchSpanProcessor,
    ConsoleSpanExporter,
)

provider = TracerProvider()
processor = BatchSpanProcessor(ConsoleSpanExporter())
provider.add_span_processor(processor)

# Sets the global default tracer provider
trace.set_tracer_provider(provider)

# Creates a tracer from the global tracer provider
tracer = trace.get_tracer("my.tracer.name")

Creating spans

To create a span, you’ll typically want it to be started as the current span.

def do_work():
    with tracer.start_as_current_span("span-name") as span:
        # do some work that 'span' will track
        print("doing some work...")
        # When the 'with' block goes out of scope, 'span' is closed for you

You can also use start_span to create a span without making it the current span. This is usually done to track concurrent or asynchronous operations.

Creating nested spans

If you have a distinct sub-operation you’d like to track as a part of another one, you can create spans to represent the relationship:

def do_work():
    with tracer.start_as_current_span("parent") as parent:
        # do some work that 'parent' tracks
        print("doing some work...")
        # Create a nested span to track nested work
        with tracer.start_as_current_span("child") as child:
            # do some work that 'child' tracks
            print("doing some nested work...")
            # the nested span is closed when it's out of scope

        # This span is also closed when it goes out of scope

When you view spans in a trace visualization tool, child will be tracked as a nested span under parent.

Creating spans with decorators

It’s common to have a single span track the execution of an entire function. In that scenario, there is a decorator you can use to reduce code:

@tracer.start_as_current_span("do_work")
def do_work():
    print("doing some work...")

Use of the decorator is equivalent to creating the span inside do_work() and ending it when do_work() is finished.

To use the decorator, you must have a tracer instance available global to your function declaration.

Get the current span

Sometimes it’s helpful to access whatever the current span is at a point in time so that you can enrich it with more information.

from opentelemetry import trace

current_span = trace.get_current_span()
# enrich 'current_span' with some information

Add attributes to a span

Attributes let you attach key/value pairs to a span so it carries more information about the current operation that it’s tracking.

from opentelemetry import trace

current_span = trace.get_current_span()

current_span.set_attribute("operation.value", 1)
current_span.set_attribute("operation.name", "Saying hello!")
current_span.set_attribute("operation.other-stuff", [1, 2, 3])

Add semantic attributes

Semantic Attributes are pre-defined Attributes that are well-known naming conventions for common kinds of data. Using Semantic Attributes lets you normalize this kind of information across your systems.

To use Semantic Attributes in Python, ensure you have the semantic conventions package:

pip install opentelemetry-semantic-conventions

Then you can use it in code:

from opentelemetry import trace
from opentelemetry.semconv.trace import SpanAttributes

// ...

current_span = trace.get_current_span()
current_span.set_attribute(SpanAttributes.HTTP_METHOD, "GET")
current_span.set_attribute(SpanAttributes.HTTP_URL, "https://opentelemetry.io/")

Adding events

An event is a human-readable message on a span that represents “something happening” during its lifetime. You can think of it as a primitive log.

from opentelemetry import trace

current_span = trace.get_current_span()

current_span.add_event("Gonna try it!")

# Do the thing

current_span.add_event("Did it!")

A span can be created with zero or more span links that causally link it to another span. A link needs a span context to be created.

from opentelemetry import trace

tracer = trace.get_tracer(__name__)

with tracer.start_as_current_span("span-1"):
    # Do something that 'span-1' tracks.
    ctx = trace.get_current_span().get_span_context()
    link_from_span_1 = trace.Link(ctx)

with tracer.start_as_current_span("span-2", links=[link_from_span_1]):
    # Do something that 'span-2' tracks.
    # The link in 'span-2' is causally associated it with the 'span-1',
    # but it is not a child span.
    pass

Set span status

A Status can be set on a Span, typically used to specify that a Span has not completed successfully - Error. By default, all spans are Unset, which means a span completed without error. The Ok status is reserved for when you need to explicitly mark a span as successful rather than stick with the default of Unset (i.e., “without error”).

The status can be set at any time before the span is finished.

from opentelemetry import trace
from opentelemetry.trace import Status, StatusCode

current_span = trace.get_current_span()

try:
    # something that might fail
except:
    current_span.set_status(Status(StatusCode.ERROR))

Record exceptions in spans

It can be a good idea to record exceptions when they happen. It’s recommended to do this in conjunction with setting span status.

from opentelemetry import trace
from opentelemetry.trace import Status, StatusCode

current_span = trace.get_current_span()

try:
    # something that might fail

# Consider catching a more specific exception in your code
except Exception as ex:
    current_span.set_status(Status(StatusCode.ERROR))
    current_span.record_exception(ex)

Change the default propagation format

By default, OpenTelemetry Python will use the following propagation formats:

  • W3C Trace Context
  • W3C Baggage

If you have a need to change the defaults, you can do so either via environment variables or in code:

Using Environment Variables

You can set the OTEL_PROPAGATORS environment variable with a comma-separated list. Accepted values are:

  • "tracecontext": W3C Trace Context
  • "baggage": W3C Baggage
  • "b3": B3 Single
  • "b3multi": B3 Multi
  • "jaeger": Jaeger
  • "xray": AWS X-Ray (third party)
  • "ottrace": OT Trace (third party)
  • "none": No automatically configured propagator.

The default configuration is equivalent to OTEL_PROPAGATORS="tracecontext,baggage".

Using SDK APIs

Alternatively, you can change the format in code.

For example, if you need to use Zipkin’s B3 propagation format instead, you can install the B3 package:

pip install opentelemetry-propagator-b3

And then set the B3 propagator in your tracing initialization code:

from opentelemetry.propagate import set_global_textmap
from opentelemetry.propagators.b3 import B3Format

set_global_textmap(B3Format())

Note that environment variables will override what’s configured in code.

Further Reading

Metrics

To start collecting metrics, you’ll need to initialize a MeterProvider and optionally set it as the global default.

from opentelemetry import metrics
from opentelemetry.sdk.metrics import MeterProvider
from opentelemetry.sdk.metrics.export import (
    ConsoleMetricExporter,
    PeriodicExportingMetricReader,
)

metric_reader = PeriodicExportingMetricReader(ConsoleMetricExporter())
provider = MeterProvider(metric_readers=[metric_reader])

# Sets the global default meter provider
metrics.set_meter_provider(provider)

# Creates a meter from the global meter provider
meter = metrics.get_meter("my.meter.name")

Creating and using synchronous instruments

Instruments are used to make measurements of your application. Synchronous instruments are used inline with application/business processing logic, like when handling a request or calling another service.

First, create your instrument. Instruments are generally created once at the module or class level and then used inline with business logic. This example uses a Counter instrument to count the number of work items completed:

work_counter = meter.create_counter(
    "work.counter", unit="1", description="Counts the amount of work done"
)

Using the Counter’s add operation, the code below increments the count by one, using the work item’s type as an attribute.

def do_work(work_item):
    # count the work being doing
    work_counter.add(1, {"work.type": work_item.work_type})
    print("doing some work...")

Creating and using asynchronous instruments

Asynchronous instruments give the user a way to register callback functions, which are invoked on demand to make measurements. This is useful to periodically measure a value that cannot be instrumented directly. Async instruments are created with zero or more callbacks which will be invoked during metric collection. Each callback accepts options from the SDK and returns its observations.

This example uses an Asynchronous Gauge instrument to report the current config version provided by a configuration server by scraping an HTTP endpoint. First, write a callback to make observations:

from typing import Iterable
from opentelemetry.metrics import CallbackOptions, Observation


def scrape_config_versions(options: CallbackOptions) -> Iterable[Observation]:
    r = requests.get(
        "http://configserver/version_metadata", timeout=options.timeout_millis / 10**3
    )
    for metadata in r.json():
        yield Observation(
            metadata["version_num"], {"config.name": metadata["version_num"]}
        )

Note that OpenTelemetry will pass options to your callback containing a timeout. Callbacks should respect this timeout to avoid blocking indefinitely. Finally, create the instrument with the callback to register it:

meter.create_observable_gauge(
    "config.version",
    callbacks=[scrape_config_versions],
    description="The active config version for each configuration",
)

Further Reading

Logs

The logs API & SDK are currently under development.

Next Steps

You’ll also want to configure an appropriate exporter to export your telemetry data to one or more telemetry backends.