SignalFx Python Google Cloud Function Wrapper.
The SignalFx Python Google Cloud Function Wrapper is a wrapper around a Google Cloud Function Python function handler, used to instrument execution of the function and send metrics and traces to SignalFx.
To install from PyPi
$ pip install signalfx_gcf
By default, this function wrapper will send to the us0 realm. If you are not in this realm you will need to set the SIGNALFX_INGEST_ENDPOINT environment variable to the correct realm ingest endpoint (https://ingest.{REALM}.signalfx.com). To determine what realm you are in, check your profile page in the SignalFx web application (click the avatar in the upper right and click My Profile).
Note: the environment variables SIGNALFX_INGEST_ENDPOINT and SIGNALFX_AUTH_TOKEN are being deprecated and will not be supported in future releases.
SIGNALFX_ACCESS_TOKEN=access token # endpoint for both metrics and tracer. Overridden by SIGNALFX_METRICS_URL # and SIGNALFX_TRACING_URL if set SIGNALFX_ENDPOINT_URL=endpoint url # optional metrics and tracing configuration SIGNALFX_METRICS_URL=ingest endpoint [ default: https://pops.signalfx.com ] SIGNALFX_SEND_TIMEOUT=timeout in seconds for sending datapoint [ default: 0.3 ] SIGNALFX_TRACING_URL=tracing endpoint [ default: https://ingest.signalfx.com/v1/trace ]
SIGNALFX_ENDPOINT_URL can be used to configure a common endpoint for metrics and
traces, as is the case when forwarding with the Smart Gateway. The path /v1/traces
will automatically be added to the endpoint for traces.
If either SIGNALFX_TRACING_URL or SIGNALFX_METRICS_URL are set, they will take precendence over SIGNALFX_ENDPOINT_URL for their respective components.
For example, if only SIGNALFX_ENDPOINT_URL is set:
SIGNALFX_ENDPOINT_URL=<gateway_address>
both metrics and traces will be sent to the gateway address.
If SIGNALFX_ENDPOINT_URL and SIGNALFX_METRICS_URL are set:
SIGNALFX_METRICS_URL=https://pops.signalfx.com SIGNALFX_ENDPOINT_URL=<gateway_address>
Traces will be sent to the gateway and metrics will go through POPS.
There are two wrappers provided.
For metrics, decorate your handler with @signalfx_gcf.emits_metrics
import signalfx_gcf
@signalfx_gcf.emits_metrics
def handler(event, context):
# your code
For tracing, use the @signalfx_gcf.is_traced decorator
import signalfx_gcf
@signalfx_gcf.is_traced
def handler(event, context):
# your code
The decorators can be used individually or together.
The Google Cloud Function wrapper sends the following metrics to SignalFx:
| Metric Name | Type | Description |
|---|---|---|
| function.invocations | Counter | Count number of Cloud Funtion invocations |
| function.cold_starts | Counter | Count number of cold starts |
| function.errors | Counter | Count number of errors from underlying Cloud Function handler |
| function.duration | Gauge | Milliseconds in execution time of underlying Cloud Function handler |
The Google Cloud Function wrapper adds the following dimensions to all data points send to SignalFx:
| Tag | Description |
|---|---|
| gcf_region | Google Cloud Function Region |
| gcf_project_id | Google Cloud Function Project ID |
| gcf_function_name | Google Cloud Function Name |
| gcf_function_version | Google Cloud Function Version |
| function_wrapper_version | SignalFx function wrapper qualifier (e.g. signalfx_gcf_0.0.2) |
| metric_source | The literal value of ‘gcf_wrapper’ |
The tracing wrapper creates a span for the wrapper handler. This span has the following tags:
| Tag | Description |
|---|---|
| gcf_region | Google Cloud Function Region |
| gcf_project_id | Google Cloud Function Project ID |
| gcf_function_name | Google Cloud Function Name |
| gcf_function_version | Google Cloud Function Version |
| function_wrapper_version | SignalFx function wrapper qualifier (e.g. signalfx_gcf_0.0.2) |
| component | The literal value of ‘python-gcf-wrapper’ |
import signalfx_gcf
# sending application_performance metric with value 100 and dimension abc:def
signalfx_gcf.send_gauge('application_performance', 100, {'abc':'def'})
# sending counter metric with no dimension
signalfx_gcf.send_counter('database_calls', 1)
Manual instrumentation can be added to trace critical parts of your handler function.
import opentracing
tracer = opentracing.tracer
def some_function():
with tracer.start_active_span("span_name", tags=tags) as scope:
# do some work
span = scope.span
span.set_tag("example_tag", "example_value")
More examples and usage information can be found in the Jaeger Python Tracer documentation.
python setup.py bdist_wheel --universal
Apache Software License v2. Copyright © 2014-2019 SignalFx