Skip to content

Commit

Permalink
Add OpenTracing shim example (open-telemetry#282)
Browse files Browse the repository at this point in the history
Co-authored-by: Diego Hurtado <[email protected]>
Co-authored-by: Christian Neumüller <[email protected]>
Co-authored-by: Mauricio Vásquez <[email protected]>
  • Loading branch information
4 people authored and toumorokoshi committed Feb 17, 2020
1 parent 9de06e5 commit 4d1ec47
Show file tree
Hide file tree
Showing 8 changed files with 203 additions and 1 deletion.
90 changes: 90 additions & 0 deletions examples/opentracing/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,90 @@
# Overview

This example shows how to use the [`opentelemetry-ext-opentracing-shim`
package](https://github.com/open-telemetry/opentelemetry-python/tree/master/ext/opentelemetry-ext-opentracing-shim)
to interact with libraries instrumented with
[`opentracing-python`](https://github.com/opentracing/opentracing-python).

The included `rediscache` library creates spans via the OpenTracing Redis
integration,
[`redis_opentracing`](https://github.com/opentracing-contrib/python-redis).
Spans are exported via the Jaeger exporter, which is attached to the
OpenTelemetry tracer.

## Installation

### Jaeger

Install and run
[Jaeger](https://www.jaegertracing.io/docs/latest/getting-started/#all-in-one).
See the [basic tracer
example](https://github.com/open-telemetry/opentelemetry-python/tree/master/examples/basic-tracer)
for more detail.

### Redis

Install Redis following the [instructions](https://redis.io/topics/quickstart).

Make sure that the Redis server is running by executing this:

```sh
$ redis-server
```

### Python Dependencies

Install the Python dependencies in [`requirements.txt`](requirements.txt):

```sh
$ pip install -r requirements.txt
```

Alternatively, you can install the Python dependencies separately:

```sh
$ pip install \
opentelemetry-api \
opentelemetry-sdk \
opentelemetry-ext-jaeger \
opentelemetry-opentracing-shim \
redis \
redis_opentracing
```

## Run the Application

The example script calculates a few Fibonacci numbers and stores the results in
Redis. The script, the `rediscache` library, and the OpenTracing Redis
integration all contribute spans to the trace.

To run the script:

```sh
$ python main.py
```

After running, you can view the generated trace in the Jaeger UI.

#### Jaeger UI

Open the Jaeger UI in your browser at
<http://localhost:16686> and view traces for the
"OpenTracing Shim Example" service.

Each `main.py` run should generate a trace, and each trace should include
multiple spans that represent calls to Redis.

<p align="center"><img src="./images/jaeger-trace-full.png?raw=true"/></p>

Note that tags and logs (OpenTracing) and attributes and events (OpenTelemetry)
from both tracing systems appear in the exported trace.

<p align="center"><img src="./images/jaeger-span-expanded.png?raw=true"/></p>

## Useful links
- For more information on OpenTelemetry, visit: <https://opentelemetry.io/>
- For more information on tracing in Python, visit: <https://github.com/open-telemetry/opentelemetry-python>

## LICENSE

Apache License 2.0
Empty file.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
46 changes: 46 additions & 0 deletions examples/opentracing/main.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,46 @@
#!/usr/bin/env python

from opentelemetry import trace
from opentelemetry.ext import opentracing_shim
from opentelemetry.ext.jaeger import JaegerSpanExporter
from opentelemetry.sdk.trace import TracerSource
from opentelemetry.sdk.trace.export import SimpleExportSpanProcessor
from rediscache import RedisCache

# Configure the tracer using the default implementation
trace.set_preferred_tracer_source_implementation(lambda T: TracerSource())
tracer_source = trace.tracer_source()

# Configure the tracer to export traces to Jaeger
jaeger_exporter = JaegerSpanExporter(
service_name="OpenTracing Shim Example",
agent_host_name="localhost",
agent_port=6831,
)
span_processor = SimpleExportSpanProcessor(jaeger_exporter)
tracer_source.add_span_processor(span_processor)

# Create an OpenTracing shim. This implements the OpenTracing tracer API, but
# forwards calls to the underlying OpenTelemetry tracer.
opentracing_tracer = opentracing_shim.create_tracer(tracer_source)

# Our example caching library expects an OpenTracing-compliant tracer.
redis_cache = RedisCache(opentracing_tracer)

# Appication code uses an OpenTelemetry Tracer as usual.
tracer = trace.tracer_source().get_tracer(__name__)


@redis_cache
def fib(number):
"""Get the Nth Fibonacci number, cache intermediate results in Redis."""
if number < 0:
raise ValueError
if number in (0, 1):
return number
return fib(number - 1) + fib(number - 2)


with tracer.start_as_current_span("Fibonacci") as span:
span.set_attribute("is_example", "yes :)")
fib(4)
61 changes: 61 additions & 0 deletions examples/opentracing/rediscache.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,61 @@
"""
This is an example of a library written to work with opentracing-python. It
provides a simple caching decorator backed by Redis, and uses the OpenTracing
Redis integration to automatically generate spans for each call to Redis.
"""

import pickle
from functools import wraps

# FIXME The pylint disablings are needed here because the code of this
# example is being executed against the tox.ini of the main
# opentelemetry-python project. Find a way to separate the two.
import redis # pylint: disable=import-error
import redis_opentracing # pylint: disable=import-error


class RedisCache:
"""Redis-backed caching decorator, using OpenTracing!
Args:
tracer: an opentracing.tracer.Tracer
"""

def __init__(self, tracer):
redis_opentracing.init_tracing(tracer)
self.tracer = tracer
self.client = redis.StrictRedis()

def __call__(self, func):
@wraps(func)
def inner(*args, **kwargs):
with self.tracer.start_active_span("Caching decorator") as scope1:

# Pickle the call args to get a canonical key. Don't do this in
# prod!
key = pickle.dumps((func.__qualname__, args, kwargs))

pval = self.client.get(key)
if pval is not None:
val = pickle.loads(pval)
scope1.span.log_kv(
{"msg": "Found cached value", "val": val}
)
return val

scope1.span.log_kv({"msg": "Cache miss, calling function"})
with self.tracer.start_active_span(
'Call "{}"'.format(func.__name__)
) as scope2:
scope2.span.set_tag("func", func.__name__)
scope2.span.set_tag("args", str(args))
scope2.span.set_tag("kwargs", str(kwargs))

val = func(*args, **kwargs)
scope2.span.set_tag("val", str(val))

# Let keys expire after 10 seconds
self.client.setex(key, 10, pickle.dumps(val))
return val

return inner
6 changes: 6 additions & 0 deletions examples/opentracing/requirements.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
opentelemetry-api
opentelemetry-sdk
opentelemetry-ext-jaeger
opentelemetry-opentracing-shim
redis
redis_opentracing
1 change: 0 additions & 1 deletion ext/opentelemetry-ext-opentracing-shim/tests/test_shim.py
Original file line number Diff line number Diff line change
Expand Up @@ -489,7 +489,6 @@ def test_inject_text_map(self):

# Verify Format.TEXT_MAP
text_map = {}

self.shim.inject(context, opentracing.Format.TEXT_MAP, text_map)
self.assertEqual(text_map[MockHTTPTextFormat.TRACE_ID_KEY], str(1220))
self.assertEqual(text_map[MockHTTPTextFormat.SPAN_ID_KEY], str(7478))
Expand Down

0 comments on commit 4d1ec47

Please sign in to comment.