Skip to content

Latest commit

 

History

History
150 lines (99 loc) · 5.04 KB

README.md

File metadata and controls

150 lines (99 loc) · 5.04 KB
Raw

Google Cloud Logging formatter for structlog

This is an opiniated package that configures structlog to output log compatible with the Google Cloud Logging log format.

The intention of this package is to be used for applications that run in Google Kubernetes Engine (GKE) or Google Cloud Function, or any other systems that know how to send logs to Google Cloud.

As such, the package is only concerned about formatting logs, where logs are expected to be written on the standard output. Sending the logs to the actual Google Logging API is supposed to be done by an external agent.

In particular, this package provides the following configuration by default:

How to use?

Install the package with pip or your favorite Python package manager:

pip install structlog-gcp

Then, configure structlog as usual, using the Structlog processors the package provides:

import structlog
import structlog_gcp

processors = structlog_gcp.build_processors()
structlog.configure(processors=processors)

Then, you can use structlog as usual:

logger = structlog.get_logger().bind(arg1="something")

logger.info("Hello world")

converted = False
try:
    int("foobar")
    converted = True
except:
    logger.exception("Something bad happens")

if not converted:
    logger.critical("This is not supposed to happen", converted=converted)

The structlog_gcp.build_processors() function constructs structlog processors to:

  • Output logs as Google Cloud Logging format using the default Python JSON serializer.
  • Carry context variables across loggers (see structlog: Context Variables)

For more advanced usage, see Advanced Configuration

Errors

Errors are automatically reported to the Google Error Reporting service.

You can configure the service name and the version used during the report with 2 different ways:

  • By default, the library assumes to run with Cloud Function environment variables configured, in particular the K_SERVICE and K_REVISION variables.

  • You can also pass the service name and revision at configuration time with:

    import structlog
import structlog_gcp

processors = structlog_gcp.build_processors(
    service="my-service",
    version="v1.2.3",
)
structlog.configure(processors=processors)

Advanced Configuration

If you need to have more control over the processors configured by the library, you can use the structlog_gcp.build_gcp_processors() builder function.

This function only configures the Google Cloud Logging-specific processors and omits all the rest.

In particular, you can use this function:

  • If you want to have more control over the processors to be configured in structlog. You can prepend or append other processors around the Google-specific ones.
  • If you want to serialize using another JSON serializer or with specific options.

For instance:

import orjson
import structlog
from structlog.processors import JSONRenderer

import structlog_gcp


def add_open_telemetry_spans(...):
    # Cf. https://www.structlog.org/en/stable/frameworks.html#opentelemetry
    ...

gcp_processors = structlog_gcp.build_gcp_processors()

# Fine-tune processors
processors = [add_open_telemetry_spans]
processors.extend(gcp_processors)
processors.append(JSONRenderer(serializer=orjson.dumps))

structlog.configure(processors=processors)

Important

structlog_gcp.build_gcp_processors() doesn't configure a renderer and you must supply a JSON renderer of your choice for the library to work correctly.

Examples

Check out the examples folder to see how it can be used.

  • How it should appear in the Google Cloud Logging log explorer:

  • How it should appear in the Google Cloud Error Reporting dashboard:

Reference