Stackdriver Logging

The Stackdriver Logging service collects and stores logs from applications and services on the Google Cloud Platform, giving you fine-grained, programmatic control over your projects' logs. You can use the Stackdriver Logging API to:

For general information about Stackdriver Logging, read Stackdriver Logging Documentation.

The goal of google-cloud is to provide an API that is comfortable to Rubyists. Your authentication credentials are detected automatically in Google Cloud Platform (GCP), including Google Compute Engine (GCE), Google Kubernetes Engine (GKE), Google App Engine (GAE), Google Cloud Functions (GCF) and Cloud Run. In other environments you can configure authentication easily, either directly in your code or via environment variables. Read more about the options for connecting in the Authentication Guide.

Listing log entries

Stackdriver Logging gathers log entries from many services, including Google App Engine and Google Compute Engine. (See the List of Log Types.) In addition, you can write your own log entries to the service.

Project#entries returns the Entry records belonging to your project:

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
entries = logging.entries
entries.each do |e|
  puts "[#{e.timestamp}] #{e.log_name} #{e.payload.inspect}"
end

You can narrow the results to a single log using an advanced logs filter. A log is a named collection of entries. Logs can be produced by Google Cloud Platform services, by third-party services, or by your applications. For example, the log compute.googleapis.com/activity_log is produced by Google Compute Engine. Logs are simply referenced by name in google-cloud. There is no Log type in google-cloud or Log resource in the Stackdriver Logging API.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
entries = logging.entries filter: "logName:syslog"
entries.each do |e|
  puts "[#{e.timestamp}] #{e.payload.inspect}"
end

You can also order the log entries by timestamp.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
entries = logging.entries order: "timestamp desc"
entries.each do |e|
  puts "[#{e.timestamp}] #{e.log_name}"
end

Exporting log entries

Stackdriver Logging lets you export log entries to destinations including Google Cloud Storage buckets (for long term log storage), Google BigQuery datasets (for log analysis), and Google Pub/Sub (for streaming to other applications).

Creating sinks

A Sink is an object that lets you to specify a set of log entries to export.

In addition to the name of the sink and the export destination, Project#create_sink accepts an advanced logs filter to narrow the collection.

Before creating the sink, ensure that you have granted cloud-logs@google.com permission to write logs to the destination. See Exporting Logs (V2).

require "google/cloud/storage"
require "google/cloud/logging"

storage = Google::Cloud::Storage.new

bucket = storage.create_bucket "my-logs-bucket"

# Grant owner permission to Stackdriver Logging service
email = "cloud-logs@google.com"
bucket.acl.add_owner "group-#{email}"

logging = Google::Cloud::Logging.new

sink = logging.create_sink "my-sink",
                           "storage.googleapis.com/#{bucket.id}"

When you create a sink, only new log entries are exported. Stackdriver Logging does not send previously-ingested log entries to the sink's destination.

Listing sinks

You can also list the sinks belonging to your project with Project#sinks.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
sinks = logging.sinks
sinks.each do |s|
  puts "#{s.name}: #{s.filter} -> #{s.destination}"
end

Creating logs-based metrics

You can use log entries in your project as the basis for Google Cloud Monitoring metrics. These metrics can then be used to produce Cloud Monitoring reports and alerts.

Creating metrics

A metric is a measured value that can be used to assess a system. Use Project#create_metric to configure a Metric based on a collection of log entries matching an advanced logs filter.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
metric = logging.create_metric "errors", "severity>=ERROR"

Listing metrics

You can also list the metrics belonging to your project with Project#metrics.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
metrics = logging.metrics
metrics.each do |m|
  puts "#{m.name}: #{m.filter}"
end

Writing log entries

An Google::Cloud::Logging::Entry is composed of metadata and a payload. The payload is traditionally a message string, but in Stackdriver Logging it can also be a JSON or protocol buffer object. A single log can have entries with different payload types. In addition to the payload, your argument(s) to Project#write_entries must also contain a log name and a resource.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new

entry = logging.entry
entry.payload = "Job started."
entry.log_name = "my_app_log"
entry.resource.type = "gae_app"
entry.resource.labels[:module_id] = "1"
entry.resource.labels[:version_id] = "20150925t173233"

logging.write_entries entry

To write a JSON payload to the log, simply pass a hash argument:

require "google/cloud/logging"

logging = Google::Cloud::Logging.new

entry = logging.entry
entry.payload = { "stats" => { "a" => 8, "b" => 12.5} }
entry.log_name = "my_app_log"
entry.resource.type = "gae_app"
entry.resource.labels[:module_id] = "1"
entry.resource.labels[:version_id] = "20150925t173233"

logging.write_entries entry

If you write a collection of log entries, you can provide the log name, resource, and/or labels hash to be used for all of the entries, and omit these values from the individual entries.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new

entry1 = logging.entry
entry1.payload = "Job started."
entry2 = logging.entry
entry2.payload = "Job completed."
labels = { job_size: "large", job_code: "red" }

resource = logging.resource "gae_app",
                            "module_id" => "1",
                            "version_id" => "20150925t173233"

logging.write_entries [entry1, entry2],
                      log_name: "my_app_log",
                      resource: resource,
                      labels: labels

Normally, writing log entries is done synchronously; the call to Project#write_entries will block until it has either completed transmitting the data or encountered an error. To "fire and forget" without blocking, use AsyncWriter; it spins up a background thread that writes log entries in batches. Calls to AsyncWriter#write_entries simply add entries to its work queue and return immediately.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new
async = logging.async_writer

entry1 = logging.entry
entry1.payload = "Job started."
entry2 = logging.entry
entry2.payload = "Job completed."
labels = { job_size: "large", job_code: "red" }

resource = logging.resource "gae_app",
                            "module_id" => "1",
                            "version_id" => "20150925t173233"

async.write_entries [entry1, entry2],
                    log_name: "my_app_log",
                    resource: resource,
                    labels: labels,
                    partial_success: true

Creating a Ruby Logger implementation

If your environment requires a logger instance that is API-compatible with Ruby's standard library Logger, you can use Project#logger to create one.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new

resource = logging.resource "gae_app",
                            module_id: "1",
                            version_id: "20150925t173233"

logger = logging.logger "my_app_log", resource, env: :production
logger.info "Job started."

By default, the logger instance writes log entries asynchronously in a background thread using an AsyncWriter. If you want to customize or disable asynchronous writing, you may call the Logger constructor directly.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new

resource = logging.resource "gae_app",
                            module_id: "1",
                            version_id: "20150925t173233"

logger = Google::Cloud::Logging::Logger.new logging,
                                            "my_app_log",
                                            resource,
                                            {env: :production}
logger.info "Log entry written synchronously."

Configuring timeout

You can configure the request timeout value in seconds.

require "google/cloud/logging"

logging = Google::Cloud::Logging.new timeout: 120

Additional information

Stackdriver Logging can be configured to be used in Rack applications or to use gRPC's logging. To learn more, see the Instrumentation Guide and Logging guide.