Metrics instrumentation guide

This guide describes how to develop Service Ping metrics using metrics instrumentation.

Nomenclature

• Instrumentation class:

  • Inherits one of the metric classes: DatabaseMetric, RedisMetric, RedisHLLMetric or GenericMetric.
  • Implements the logic that calculates the value for a Service Ping metric.

• Metric definition The Service Data metric YAML definition.

• Hardening: Hardening a method is the process that ensures the method fails safe, returning a fallback value like -1.

How it works

A metric definition has the instrumentation_class field, which can be set to a class.

The defined instrumentation class should have one of the existing metric classes: DatabaseMetric, RedisMetric, RedisHLLMetric, or GenericMetric.

Using the instrumentation classes ensures that metrics can fail safe individually, without breaking the entire process of Service Ping generation.

We have built a domain-specific language (DSL) to define the metrics instrumentation.

Database metrics

• operation: Operations for the given relation, one of count, distinct_count.
• relation: ActiveRecord::Relation for the objects we want to perform the operation.
• start: Specifies the start value of the batch counting, by default is relation.minimum(:id).
• finish: Specifies the end value of the batch counting, by default is relation.maximum(:id).
• cache_start_and_finish_as: Specifies the cache key for start and finish values and sets up caching them. Use this call when start and finish are expensive queries that should be reused between different metric calculations.

Example of a merge request that adds a database metric.

module Gitlab
  module Usage
    module Metrics
      module Instrumentations
        class CountBoardsMetric < DatabaseMetric
          operation :count

          relation { Board }
        end
      end
    end
  end
end

Redis metrics

Example of a merge request that adds a Redis metric.

Count unique values for source_code_pushes event.

Required options:

• event: the event name.
• counter_class: one of the counter classes from the Gitlab::UsageDataCounters namespace; it should implement read method or inherit it from BaseCounter.

time_frame: all
data_source: redis
instrumentation_class: 'RedisMetric'
options:
  event: pushes
  counter_class: SourceCodeCounter

Redis HyperLogLog metrics

Example of a merge request that adds a RedisHLL metric.

Count unique values for i_quickactions_approve event.

time_frame: 28d
data_source: redis_hll
instrumentation_class: 'RedisHLLMetric'
options:
  events:
    - i_quickactions_approve

Generic metrics

Example of a merge request that adds a generic metric.

module Gitlab
  module Usage
    module Metrics
      module Instrumentations
        class UuidMetric < GenericMetric
          value do
            Gitlab::CurrentSettings.uuid
          end
        end
      end
    end
  end
end

Support for instrumentation classes

There is support for:

• count, distinct_count, estimate_batch_distinct_count for database metrics.
• Redis metrics.
• Redis HLL metrics.
• Generic metrics, which are metrics based on settings or configurations.

There is no support for:

• add, sum, histogram for database metrics.

You can track the progress to support these.

Create a new metric instrumentation class

To create a stub instrumentation for a Service Ping metric, you can use a dedicated generator:

The generator takes the class name as an argument and the following options:

• --type=TYPE Required. Indicates the metric type. It must be one of: database, generic, redis.
• --operation Required for database type. It must be one of: count, distinct_count, estimate_batch_distinct_count.
• --ee Indicates if the metric is for EE.

rails generate gitlab:usage_metric CountIssues --type database
        create lib/gitlab/usage/metrics/instrumentations/count_issues_metric.rb
        create spec/lib/gitlab/usage/metrics/instrumentations/count_issues_metric_spec.rb