Merge branch 'alerts-for-built-in-metrics' into 'master'

Import common metrics into database.

Closes gitlab-ee#6948

See merge request gitlab-org/gitlab-ce!21459

app/models/project.rb

@@ -232,6 +232,8 @@ class Project < ActiveRecord::Base
   has_many :clusters, through: :cluster_project, class_name: 'Clusters::Cluster'
   has_many :cluster_ingresses, through: :clusters, source: :application_ingress, class_name: 'Clusters::Applications::Ingress'
 
+  has_many :prometheus_metrics
+
   # Container repositories need to remove data from the container registry,
   # which is not managed by the DB. Hence we're still using dependent: :destroy
   # here.

app/models/prometheus_metric.rb

+# frozen_string_literal: true
+
+class PrometheusMetric < ActiveRecord::Base
+  belongs_to :project, validate: true, inverse_of: :prometheus_metrics
+
+  enum group: {
+    # built-in groups
+    nginx_ingress: -1,
+    ha_proxy: -2,
+    aws_elb: -3,
+    nginx: -4,
+    kubernetes: -5,
+
+    # custom/user groups
+    business: 0,
+    response: 1,
+    system: 2
+  }
+
+  validates :title, presence: true
+  validates :query, presence: true
+  validates :group, presence: true
+  validates :y_label, presence: true
+  validates :unit, presence: true
+
+  validates :project, presence: true, unless: :common?
+  validates :project, absence: true, if: :common?
+
+  scope :common, -> { where(common: true) }
+
+  GROUP_TITLES = {
+    # built-in groups
+    nginx_ingress: _('Response metrics (NGINX Ingress)'),
+    ha_proxy: _('Response metrics (HA Proxy)'),
+    aws_elb: _('Response metrics (AWS ELB)'),
+    nginx: _('Response metrics (NGINX)'),
+    kubernetes: _('System metrics (Kubernetes)'),
+
+    # custom/user groups
+    business: _('Business metrics (Custom)'),
+    response: _('Response metrics (Custom)'),
+    system: _('System metrics (Custom)')
+  }.freeze
+
+  REQUIRED_METRICS = {
+    nginx_ingress: %w(nginx_upstream_responses_total nginx_upstream_response_msecs_avg),
+    ha_proxy: %w(haproxy_frontend_http_requests_total haproxy_frontend_http_responses_total),
+    aws_elb: %w(aws_elb_request_count_sum aws_elb_latency_average aws_elb_httpcode_backend_5_xx_sum),
+    nginx: %w(nginx_server_requests nginx_server_requestMsec),
+    kubernetes: %w(container_memory_usage_bytes container_cpu_usage_seconds_total)
+  }.freeze
+
+  def group_title
+    GROUP_TITLES[group.to_sym]
+  end
+
+  def required_metrics
+    REQUIRED_METRICS[group.to_sym].to_a.map(&:to_s)
+  end
+
+  def to_query_metric
+    Gitlab::Prometheus::Metric.new(id: id, title: title, required_metrics: required_metrics, weight: 0, y_label: y_label, queries: queries)
+  end
+
+  def queries
+    [
+      {
+        query_range: query,
+        unit: unit,
+        label: legend,
+        series: query_series
+      }.compact
+    ]
+  end
+
+  def query_series
+    case legend
+    when 'Status Code'
+      [{
+        label: 'status_code',
+        when: [
+          { value: '2xx', color: 'green' },
+          { value: '4xx', color: 'orange' },
+          { value: '5xx', color: 'red' }
+        ]
+      }]
+    end
+  end
+end

changelogs/unreleased/import-all-common-metrics-into-database.yml

+---
+title: Import all common metrics into database
+merge_request: 21459
+author:
+type: changed

config/prometheus/additional_metrics.yml → config/prometheus/common_metrics.yml

@@ -7,7 +7,8 @@
       - nginx_upstream_responses_total
     weight: 1
     queries:
-    - query_range: 'sum(rate(nginx_upstream_responses_total{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) by (status_code)'
+    - id: response_metrics_nginx_ingress_throughput_status_code
+      query_range: 'sum(rate(nginx_upstream_responses_total{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) by (status_code)'
       unit: req / sec
       label: Status Code
       series:
@@ -25,7 +26,8 @@
       - nginx_upstream_response_msecs_avg
     weight: 1
     queries:
-    - query_range: 'avg(nginx_upstream_response_msecs_avg{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"})'
+    - id: response_metrics_nginx_ingress_latency_pod_average
+      query_range: 'avg(nginx_upstream_response_msecs_avg{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"})'
       label: Pod average
       unit: ms
   - title: "HTTP Error Rate"
@@ -34,7 +36,8 @@
       - nginx_upstream_responses_total
     weight: 1
     queries:
-    - query_range: 'sum(rate(nginx_upstream_responses_total{status_code="5xx", upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) / sum(rate(nginx_upstream_responses_total{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) * 100'
+    - id: response_metrics_nginx_ingress_http_error_rate
+      query_range: 'sum(rate(nginx_upstream_responses_total{status_code="5xx", upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) / sum(rate(nginx_upstream_responses_total{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) * 100'
       label: 5xx Errors
       unit: "%"
 - group: Response metrics (HA Proxy)
@@ -46,10 +49,12 @@
       - haproxy_frontend_http_requests_total
     weight: 1
     queries:
-    - query_range: 'sum(rate(haproxy_frontend_http_requests_total{%{environment_filter}}[2m])) by (code)'
+    - id: response_metrics_ha_proxy_throughput_status_code
+      query_range: 'sum(rate(haproxy_frontend_http_requests_total{%{environment_filter}}[2m])) by (code)'
       unit: req / sec
+      label: Status Code
       series:
-        - label: code
+        - label: status_code
           when:
             - value: 2xx
               color: green
@@ -63,7 +68,8 @@
       - haproxy_frontend_http_responses_total
     weight: 1
     queries:
-    - query_range: 'sum(rate(haproxy_frontend_http_responses_total{code="5xx",%{environment_filter}}[2m])) / sum(rate(haproxy_frontend_http_responses_total{%{environment_filter}}[2m]))'
+    - id: response_metrics_ha_proxy_http_error_rate
+      query_range: 'sum(rate(haproxy_frontend_http_responses_total{code="5xx",%{environment_filter}}[2m])) / sum(rate(haproxy_frontend_http_responses_total{%{environment_filter}}[2m]))'
       label: HTTP Errors
       unit: "%"
 - group: Response metrics (AWS ELB)
@@ -75,7 +81,8 @@
       - aws_elb_request_count_sum
     weight: 1
     queries:
-    - query_range: 'sum(aws_elb_request_count_sum{%{environment_filter}}) / 60'
+    - id: response_metrics_aws_elb_throughput_requests
+      query_range: 'sum(aws_elb_request_count_sum{%{environment_filter}}) / 60'
       label: Total
       unit: req / sec
   - title: "Latency"
@@ -84,7 +91,8 @@
       - aws_elb_latency_average
     weight: 1
     queries:
-    - query_range: 'avg(aws_elb_latency_average{%{environment_filter}}) * 1000'
+    - id: response_metrics_aws_elb_latency_average
+      query_range: 'avg(aws_elb_latency_average{%{environment_filter}}) * 1000'
       label: Average
       unit: ms
   - title: "HTTP Error Rate"
@@ -94,7 +102,8 @@
       - aws_elb_httpcode_backend_5_xx_sum
     weight: 1
     queries:
-    - query_range: 'sum(aws_elb_httpcode_backend_5_xx_sum{%{environment_filter}}) / sum(aws_elb_request_count_sum{%{environment_filter}})'
+    - id: response_metrics_aws_elb_http_error_rate
+      query_range: 'sum(aws_elb_httpcode_backend_5_xx_sum{%{environment_filter}}) / sum(aws_elb_request_count_sum{%{environment_filter}})'
       label: HTTP Errors
       unit: "%"
 - group: Response metrics (NGINX)
@@ -106,7 +115,8 @@
       - nginx_server_requests
     weight: 1
     queries:
-    - query_range: 'sum(rate(nginx_server_requests{server_zone!="*", server_zone!="_", %{environment_filter}}[2m])) by (code)'
+    - id: response_metrics_nginx_throughput_status_code
+      query_range: 'sum(rate(nginx_server_requests{server_zone!="*", server_zone!="_", %{environment_filter}}[2m])) by (code)'
       unit: req / sec
       label: Status Code
       series:
@@ -124,7 +134,8 @@
       - nginx_server_requestMsec
     weight: 1
     queries:
-    - query_range: 'avg(nginx_server_requestMsec{%{environment_filter}})'
+    - id: response_metrics_nginx_latency
+      query_range: 'avg(nginx_server_requestMsec{%{environment_filter}})'
       label: Upstream
       unit: ms
   - title: "HTTP Error Rate"
@@ -133,7 +144,8 @@
       - nginx_server_requests
     weight: 1
     queries:
-    - query_range: 'sum(rate(nginx_server_requests{code="5xx", %{environment_filter}}[2m]))'
+    - id: response_metrics_nginx_http_error_rate
+      query_range: 'sum(rate(nginx_server_requests{code="5xx", %{environment_filter}}[2m]))'
       label: HTTP Errors
       unit: "errors / sec"
 - group: System metrics (Kubernetes)
@@ -145,7 +157,8 @@
       - container_memory_usage_bytes
     weight: 4
     queries:
-    - query_range: 'avg(sum(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}) by (job)) without (job)  /1024/1024/1024'
+    - id: system_metrics_kubernetes_container_memory_total
+      query_range: 'avg(sum(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}) by (job)) without (job)  /1024/1024/1024'
       label: Total
       unit: GB
   - title: "Core Usage (Total)"
@@ -154,7 +167,8 @@
      - container_cpu_usage_seconds_total
     weight: 3
     queries:
-    - query_range: 'avg(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}[15m])) by (job)) without (job)'
+    - id: system_metrics_kubernetes_container_cores_total
+      query_range: 'avg(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}[15m])) by (job)) without (job)'
       label: Total
       unit: "cores"
   - title: "Memory Usage (Pod average)"
@@ -163,15 +177,39 @@
       - container_memory_usage_bytes
     weight: 2
     queries:
-    - query_range: 'avg(sum(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}) by (job)) without (job) / count(avg(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}) without (job)) /1024/1024'
+    - id: system_metrics_kubernetes_container_memory_average
+      query_range: 'avg(sum(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}) by (job)) without (job) / count(avg(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}) without (job)) /1024/1024'
+      label: Pod average
+      unit: MB
+  - title: "Canary: Memory Usage (Pod Average)"
+    y_label: "Memory Used per Pod"
+    required_metrics:
+      - container_memory_usage_bytes
+    weight: 2
+    queries:
+    - id: system_metrics_kubernetes_container_memory_average_canary
+      query_range: 'avg(sum(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-canary-(.*)",namespace="%{kube_namespace}"}) by (job)) without (job) / count(avg(container_memory_usage_bytes{container_name!="POD",pod_name=~"^%{ci_environment_slug}-canary-(.*)",namespace="%{kube_namespace}"}) without (job)) /1024/1024'
       label: Pod average
       unit: MB
-  - title: "Core Usage (Pod average)"
+      track: canary
+  - title: "Core Usage (Pod Average)"
     y_label: "Cores per Pod"
     required_metrics:
      - container_cpu_usage_seconds_total
     weight: 1
     queries:
-    - query_range: 'avg(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}[15m])) by (job)) without (job) / count(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}[15m])) by (pod_name))'
+    - id: system_metrics_kubernetes_container_core_usage
+      query_range: 'avg(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}[15m])) by (job)) without (job) / count(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-(.*)",namespace="%{kube_namespace}"}[15m])) by (pod_name))'
       label: Pod average
-      unit: "cores"
+      unit: "cores"
\ No newline at end of file
+  - title: "Canary: Core Usage (Pod Average)"
+    y_label: "Cores per Pod"
+    required_metrics:
+     - container_cpu_usage_seconds_total
+    weight: 1
+    queries:
+    - id: system_metrics_kubernetes_container_core_usage_canary
+      query_range: 'avg(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-canary-(.*)",namespace="%{kube_namespace}"}[15m])) by (job)) without (job) / count(sum(rate(container_cpu_usage_seconds_total{container_name!="POD",pod_name=~"^%{ci_environment_slug}-canary-(.*)",namespace="%{kube_namespace}"}[15m])) by (pod_name))'
+      label: Pod average
+      unit: "cores"
+      track: canary

db/fixtures/development/99_common_metrics.rb

+# frozen_string_literal: true
+
+require Rails.root.join('db/importers/common_metrics_importer.rb')
+
+::Importers::CommonMetricsImporter.new.execute

db/fixtures/production/999_common_metrics.rb

+# frozen_string_literal: true
+
+require Rails.root.join('db/importers/common_metrics_importer.rb')
+
+::Importers::CommonMetricsImporter.new.execute

db/importers/common_metrics_importer.rb

+# frozen_string_literal: true
+
+module Importers
+  class PrometheusMetric < ActiveRecord::Base
+    enum group: {
+      # built-in groups
+      nginx_ingress: -1,
+      ha_proxy: -2,
+      aws_elb: -3,
+      nginx: -4,
+      kubernetes: -5,
+
+      # custom groups
+      business: 0,
+      response: 1,
+      system: 2
+    }
+
+    scope :common, -> { where(common: true) }
+
+    GROUP_TITLES = {
+      business: _('Business metrics (Custom)'),
+      response: _('Response metrics (Custom)'),
+      system: _('System metrics (Custom)'),
+      nginx_ingress: _('Response metrics (NGINX Ingress)'),
+      ha_proxy: _('Response metrics (HA Proxy)'),
+      aws_elb: _('Response metrics (AWS ELB)'),
+      nginx: _('Response metrics (NGINX)'),
+      kubernetes: _('System metrics (Kubernetes)')
+    }.freeze
+  end
+
+  class CommonMetricsImporter
+    MissingQueryId = Class.new(StandardError)
+
+    attr_reader :content
+
+    def initialize(file = 'config/prometheus/common_metrics.yml')
+      @content = YAML.load_file(file)
+    end
+
+    def execute
+      process_content do |id, attributes|
+        find_or_build_metric!(id)
+          .update!(**attributes)
+      end
+    end
+
+    private
+
+    def process_content(&blk)
+      content.map do |group|
+        process_group(group, &blk)
+      end
+    end
+
+    def process_group(group, &blk)
+      attributes = {
+        group: find_group_title_key(group['group'])
+      }
+
+      group['metrics'].map do |metric|
+        process_metric(metric, attributes, &blk)
+      end
+    end
+
+    def process_metric(metric, attributes, &blk)
+      attributes = attributes.merge(
+        title: metric['title'],
+        y_label: metric['y_label'])
+
+      metric['queries'].map do |query|
+        process_metric_query(query, attributes, &blk)
+      end
+    end
+
+    def process_metric_query(query, attributes, &blk)
+      attributes = attributes.merge(
+        legend: query['label'],
+        query: query['query_range'],
+        unit: query['unit'])
+
+      yield(query['id'], attributes)
+    end
+
+    def find_or_build_metric!(id)
+      raise MissingQueryId unless id
+
+      PrometheusMetric.common.find_by(identifier: id) ||
+        PrometheusMetric.new(common: true, identifier: id)
+    end
+
+    def find_group_title_key(title)
+      PrometheusMetric.groups[find_group_title(title)]
+    end
+
+    def find_group_title(title)
+      PrometheusMetric::GROUP_TITLES.invert[title]
+    end
+  end
+end

db/migrate/20180101160629_create_prometheus_metrics.rb

+# frozen_string_literal: true
+
+class CreatePrometheusMetrics < ActiveRecord::Migration
+  DOWNTIME = false
+
+  def change
+    create_table :prometheus_metrics do |t|
+      t.references :project, index: true, foreign_key: { on_delete: :cascade }, null: false
+      t.string :title, null: false
+      t.string :query, null: false
+      t.string :y_label
+      t.string :unit
+      t.string :legend
+      t.integer :group, null: false, index: true
+      t.timestamps_with_timezone null: false
+    end
+  end
+end

db/migrate/20180101160630_change_project_id_for_prometheus_metrics.rb

+# frozen_string_literal: true
+
+class ChangeProjectIdForPrometheusMetrics < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+
+  def change
+    change_column_null :prometheus_metrics, :project_id, true
+  end
+end

db/migrate/20180831164904_fix_prometheus_metric_query_limits.rb

+# frozen_string_literal: true
+
+# See http://doc.gitlab.com/ce/development/migration_style_guide.html
+# for more information on how to write migrations for GitLab.
+require Rails.root.join('db/migrate/prometheus_metrics_limits_to_mysql')
+
+class FixPrometheusMetricQueryLimits < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+
+  def up
+    PrometheusMetricsLimitsToMysql.new.up
+  end
+
+  def down
+    # no-op
+  end
+end

db/migrate/20180831164905_add_common_to_prometheus_metrics.rb

+# frozen_string_literal: true
+
+class AddCommonToPrometheusMetrics < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+
+  disable_ddl_transaction!
+
+  def up
+    add_column_with_default(:prometheus_metrics, :common, :boolean, default: false)
+  end
+
+  def down
+    remove_column(:prometheus_metrics, :common)
+  end
+end

db/migrate/20180831164907_add_index_on_common_for_prometheus_metrics.rb

+# frozen_string_literal: true
+
+class AddIndexOnCommonForPrometheusMetrics < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+
+  disable_ddl_transaction!
+
+  def up
+    add_concurrent_index :prometheus_metrics, :common
+  end
+
+  def down
+    remove_concurrent_index :prometheus_metrics, :common
+  end
+end

db/migrate/20180831164908_add_identifier_to_prometheus_metric.rb

+# frozen_string_literal: true
+
+class AddIdentifierToPrometheusMetric < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+
+  def change
+    add_column :prometheus_metrics, :identifier, :string
+  end
+end

db/migrate/20180831164909_add_index_for_identifier_to_prometheus_metric.rb

+# frozen_string_literal: true
+
+class AddIndexForIdentifierToPrometheusMetric < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+
+  disable_ddl_transaction!
+
+  def up
+    add_concurrent_index :prometheus_metrics, :identifier, unique: true
+  end
+
+  def down
+    remove_concurrent_index :prometheus_metrics, :identifier, unique: true
+  end
+end

db/migrate/20180831164910_import_common_metrics.rb

+# frozen_string_literal: true
+
+class ImportCommonMetrics < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  require Rails.root.join('db/importers/common_metrics_importer.rb')
+
+  DOWNTIME = false
+
+  def up
+    Importers::CommonMetricsImporter.new.execute
+  end
+
+  def down
+    # no-op
+  end
+end

db/migrate/prometheus_metrics_limits_to_mysql.rb

+class PrometheusMetricsLimitsToMysql < ActiveRecord::Migration
+  DOWNTIME = false
+
+  def up
+    return unless Gitlab::Database.mysql?
+
+    change_column :prometheus_metrics, :query, :text, limit: 4096, default: nil
+  end
+
+  def down
+  end
+end

db/schema.rb

@@ -1700,6 +1700,25 @@ ActiveRecord::Schema.define(version: 20180901171833) do
   add_index "projects", ["star_count"], name: "index_projects_on_star_count", using: :btree
   add_index "projects", ["visibility_level"], name: "index_projects_on_visibility_level", using: :btree
 
+  create_table "prometheus_metrics", force: :cascade do |t|
+    t.integer "project_id"
+    t.string "title", null: false
+    t.string "query", null: false
+    t.string "y_label"
+    t.string "unit"
+    t.string "legend"
+    t.integer "group", null: false
+    t.datetime_with_timezone "created_at", null: false
+    t.datetime_with_timezone "updated_at", null: false
+    t.boolean "common", default: false, null: false
+    t.string "identifier"
+  end
+
+  add_index "prometheus_metrics", ["common"], name: "index_prometheus_metrics_on_common", using: :btree
+  add_index "prometheus_metrics", ["group"], name: "index_prometheus_metrics_on_group", using: :btree
+  add_index "prometheus_metrics", ["identifier"], name: "index_prometheus_metrics_on_identifier", unique: true, using: :btree
+  add_index "prometheus_metrics", ["project_id"], name: "index_prometheus_metrics_on_project_id", using: :btree
+
   create_table "protected_branch_merge_access_levels", force: :cascade do |t|
     t.integer "protected_branch_id", null: false
     t.integer "access_level", default: 40, null: false
@@ -2380,6 +2399,7 @@ ActiveRecord::Schema.define(version: 20180901171833) do
   add_foreign_key "project_import_data", "projects", name: "fk_ffb9ee3a10", on_delete: :cascade
   add_foreign_key "project_mirror_data", "projects", on_delete: :cascade
   add_foreign_key "project_statistics", "projects", on_delete: :cascade
+  add_foreign_key "prometheus_metrics", "projects", on_delete: :cascade
   add_foreign_key "protected_branch_merge_access_levels", "protected_branches", name: "fk_8a3072ccb3", on_delete: :cascade
   add_foreign_key "protected_branch_push_access_levels", "protected_branches", name: "fk_9ffc86a3d9", on_delete: :cascade
   add_foreign_key "protected_branches", "projects", name: "fk_7a9c6d93e7", on_delete: :cascade

doc/development/README.md

@@ -47,6 +47,7 @@ description: 'Learn how to contribute to GitLab.'
 - [How to dump production data to staging](db_dump.md)
 - [Working with the GitHub importer](github_importer.md)
 - [Working with Merge Request diffs](diffs.md)
+- [Prometheus metrics](prometheus_metrics.md)
 
 ## Performance guides
 

doc/development/prometheus_metrics.md

+# Working with Prometheus Metrics
+
+## Adding to the library
+
+We strive to support the 2-4 most important metrics for each common system service that supports Prometheus. If you are looking for support for a particular exporter which has not yet been added to the library, additions can be made [to the `common_metrics.yml`](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/config/prometheus/common_metrics.yml) file.
+
+### Query identifier
+
+The requirement for adding a new metric is to make each query to have an unique identifier which is used to update the metric later when changed:
+
+```yaml
+- group: Response metrics (NGINX Ingress)
+  metrics:
+  - title: "Throughput"
+    y_label: "Requests / Sec"
+    queries:
+    - id: response_metrics_nginx_ingress_throughput_status_code
+      query_range: 'sum(rate(nginx_upstream_responses_total{upstream=~"%{kube_namespace}-%{ci_environment_slug}-.*"}[2m])) by (status_code)'
+      unit: req / sec
+      label: Status Code
+```
+
+### Update existing metrics
+
+After you add or change existing _common_ metric you have to create a new database migration that will query and update all existing metrics.
+
+NOTE: **Note:**
+If a query metric (which is identified by `id:`) is removed it will not be removed from database by default.
+You might want to add additional database migration that makes a decision what to do with removed one.
+For example: you might be interested in migrating all dependent data to a different metric.
+
+```ruby
+class ImportCommonMetrics < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  require Rails.root.join('db/importers/common_metrics_importer.rb')
+
+  DOWNTIME = false
+
+  def up
+    Importers::CommonMetricsImporter.new.execute
+  end
+
+  def down
+    # no-op
+  end
+end
+```

doc/user/project/integrations/prometheus_library/metrics.md

@@ -17,9 +17,3 @@ GitLab retrieves performance data from the configured Prometheus server, and att
 
 In order to isolate and only display relevant metrics for a given environment, GitLab needs a method to detect which labels are associated. To do that,
 GitLab uses the defined queries and fills in the environment specific variables. Typically this involves looking for the [$CI_ENVIRONMENT_SLUG](../../../../ci/variables/README.md#predefined-variables-environment-variables), but may also include other information such as the project's Kubernetes namespace. Each search query is defined in the [exporter specific documentation](#prometheus-metrics-library).
-
-## Adding to the library
-
-We strive to support the 2-4 most important metrics for each common system service that supports Prometheus. If you are looking for support for a particular exporter which has not yet been added to the library, additions can be made [to the `additional_metrics.yml`](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/config/prometheus/additional_metrics.yml) file.
-
-> Note: The library is only for monitoring public, common, system services which all customers can benefit from. Support for monitoring [customer proprietary metrics](https://gitlab.com/gitlab-org/gitlab-ee/issues/2273) will be added in a subsequent release.

lib/gitlab/import_export/import_export.yml

@@ -64,6 +64,7 @@ project_tree:
     - :create_access_levels
   - :project_feature
   - :custom_attributes
+  - :prometheus_metrics
   - :project_badges
   - :ci_cd_settings
 
@@ -108,6 +109,9 @@ excluded_attributes:
     - :remote_mirror_available_overridden
     - :description_html
     - :repository_languages
+  prometheus_metrics:
+    - :common
+    - :identifier
   snippets:
     - :expired_at
   merge_request_diff:

lib/gitlab/prometheus/additional_metrics_parser.rb

@@ -5,7 +5,7 @@ module Gitlab
       MUTEX = Mutex.new
       extend self
 
-      def load_groups_from_yaml(file_name = 'additional_metrics.yml')
+      def load_groups_from_yaml(file_name)
         yaml_metrics_raw(file_name).map(&method(:group_from_entry))
       end
 

lib/gitlab/prometheus/metric_group.rb

@@ -4,10 +4,13 @@ module Gitlab
       include ActiveModel::Model
 
       attr_accessor :name, :priority, :metrics
+
       validates :name, :priority, :metrics, presence: true
 
       def self.common_metrics
-        AdditionalMetricsParser.load_groups_from_yaml
+        ::PrometheusMetric.common.group_by(&:group_title).map do |name, metrics|
+          MetricGroup.new(name: name, priority: 0, metrics: metrics.map(&:to_query_metric))
+        end
       end
 
       # EE only

lib/tasks/migrate/add_limits_mysql.rake

@@ -3,6 +3,7 @@ require Rails.root.join('db/migrate/markdown_cache_limits_to_mysql')
 require Rails.root.join('db/migrate/merge_request_diff_file_limits_to_mysql')
 require Rails.root.join('db/migrate/limits_ci_build_trace_chunks_raw_data_for_mysql')
 require Rails.root.join('db/migrate/gpg_keys_limits_to_mysql')
+require Rails.root.join('db/migrate/prometheus_metrics_limits_to_mysql')
 
 desc "GitLab | Add limits to strings in mysql database"
 task add_limits_mysql: :environment do
@@ -12,4 +13,5 @@ task add_limits_mysql: :environment do
   MergeRequestDiffFileLimitsToMysql.new.up
   LimitsCiBuildTraceChunksRawDataForMysql.new.up
   IncreaseMysqlTextLimitForGpgKeys.new.up
+  PrometheusMetricsLimitsToMysql.new.up
 end

locale/gitlab.pot

@@ -1035,6 +1035,9 @@ msgstr ""
 msgid "Browse files"
 msgstr ""
 
+msgid "Business metrics (Custom)"
+msgstr ""
+
 msgid "ByAuthor|by"
 msgstr ""
 
@@ -4917,6 +4920,21 @@ msgstr ""
 msgid "Resolve discussion"
 msgstr ""
 
+msgid "Response metrics (AWS ELB)"
+msgstr ""
+
+msgid "Response metrics (Custom)"
+msgstr ""
+
+msgid "Response metrics (HA Proxy)"
+msgstr ""
+
+msgid "Response metrics (NGINX Ingress)"
+msgstr ""
+
+msgid "Response metrics (NGINX)"
+msgstr ""
+
 msgid "Resume"
 msgstr ""
 
@@ -5500,6 +5518,12 @@ msgstr ""
 msgid "System Info"
 msgstr ""
 
+msgid "System metrics (Custom)"
+msgstr ""
+
+msgid "System metrics (Kubernetes)"
+msgstr ""
+
 msgid "Tag (%{tag_count})"
 msgid_plural "Tags (%{tag_count})"
 msgstr[0] ""

spec/db/development/import_common_metrics_spec.rb

+# frozen_string_literal: true
+
+require 'spec_helper'
+
+describe 'Import metrics on development seed' do
+  subject { load Rails.root.join('db', 'fixtures', 'development', '99_common_metrics.rb') }
+
+  it "imports all prometheus metrics" do
+    expect(PrometheusMetric.common).to be_empty
+
+    subject
+
+    expect(PrometheusMetric.common).not_to be_empty
+  end
+end

spec/db/importers/common_metrics_importer_spec.rb

+# frozen_string_literal: true
+
+require 'rails_helper'
+require Rails.root.join("db", "importers", "common_metrics_importer.rb")
+
+describe Importers::PrometheusMetric do
+  it 'group enum equals ::PrometheusMetric' do
+    expect(described_class.groups).to eq(::PrometheusMetric.groups)
+  end
+
+  it 'GROUP_TITLES equals ::PrometheusMetric' do
+    expect(described_class::GROUP_TITLES).to eq(::PrometheusMetric::GROUP_TITLES)
+  end
+end
+
+describe Importers::CommonMetricsImporter do
+  subject { described_class.new }
+
+  context "does import common_metrics.yml" do
+    let(:groups) { subject.content }
+    let(:metrics) { groups.map { |group| group['metrics'] }.flatten }
+    let(:queries) { metrics.map { |group| group['queries'] }.flatten }
+    let(:query_ids) { queries.map { |query| query['id'] } }
+
+    before do
+      subject.execute
+    end
+
+    it "has the same amount of groups" do
+      expect(PrometheusMetric.common.group(:group).count.count).to eq(groups.count)
+    end
+
+    it "has the same amount of metrics" do
+      expect(PrometheusMetric.common.group(:group, :title).count.count).to eq(metrics.count)
+    end
+
+    it "has the same amount of queries" do
+      expect(PrometheusMetric.common.count).to eq(queries.count)
+    end
+
+    it "does not have duplicate IDs" do
+      expect(query_ids).to eq(query_ids.uniq)
+    end
+
+    it "imports all IDs" do
+      expect(PrometheusMetric.common.pluck(:identifier)).to contain_exactly(*query_ids)
+    end
+  end
+
+  context 'does import properly all fields' do
+    let(:query_identifier) { 'response-metric' }
+    let(:group) do
+      {
+        group: 'Response metrics (NGINX Ingress)',
+        metrics: [{
+          title: "Throughput",
+          y_label: "Requests / Sec",
+          queries: [{
+            id: query_identifier,
+            query_range: 'my-query',
+            unit: 'my-unit',
+            label: 'status code'
+          }]
+        }]
+      }
+    end
+
+    before do
+      expect(subject).to receive(:content) { [group.deep_stringify_keys] }
+    end
+
+    shared_examples 'stores metric' do
+      let(:metric) { PrometheusMetric.find_by(identifier: query_identifier) }
+
+      it 'with all data' do
+        expect(metric.group).to eq('nginx_ingress')
+        expect(metric.title).to eq('Throughput')
+        expect(metric.y_label).to eq('Requests / Sec')
+        expect(metric.unit).to eq('my-unit')
+        expect(metric.legend).to eq('status code')
+        expect(metric.query).to eq('my-query')
+      end
+    end
+
+    context 'if ID is missing' do
+      let(:query_identifier) { }
+
+      it 'raises exception' do
+        expect { subject.execute }.to raise_error(described_class::MissingQueryId)
+      end
+    end
+
+    context 'for existing common metric with different ID' do
+      let!(:existing_metric) { create(:prometheus_metric, :common, identifier: 'my-existing-metric') }
+
+      before do
+        subject.execute
+      end
+
+      it_behaves_like 'stores metric' do
+        it 'and existing metric is not changed' do
+          expect(metric).not_to eq(existing_metric)
+        end
+      end
+    end
+
+    context 'when metric with ID exists ' do
+      let!(:existing_metric) { create(:prometheus_metric, :common, identifier: 'response-metric') }
+
+      before do
+        subject.execute
+      end
+
+      it_behaves_like 'stores metric' do
+        it 'and existing metric is changed' do
+          expect(metric).to eq(existing_metric)
+        end
+      end
+    end
+  end
+end

spec/db/production/import_common_metrics_spec.rb

+# frozen_string_literal: true
+
+require 'spec_helper'
+
+describe 'Import metrics on production seed' do
+  subject { load Rails.root.join('db', 'fixtures', 'production', '999_common_metrics.rb') }
+
+  it "imports all prometheus metrics" do
+    expect(PrometheusMetric.common).to be_empty
+
+    subject
+
+    expect(PrometheusMetric.common).not_to be_empty
+  end
+end

spec/factories/prometheus_metrics.rb

+# frozen_string_literal: true
+
+FactoryBot.define do
+  factory :prometheus_metric, class: PrometheusMetric do
+    title 'title'
+    query 'avg(metric)'
+    y_label 'y_label'
+    unit 'm/s'
+    group :business
+    project
+    legend 'legend'
+
+    trait :common do
+      common true
+      project nil
+    end
+  end
+end

spec/lib/gitlab/import_export/all_models.yml

@@ -288,6 +288,7 @@ project:
 - fork_network_member
 - fork_network
 - custom_attributes
+- prometheus_metrics
 - lfs_file_locks
 - project_badges
 - source_of_merge_requests
@@ -303,6 +304,8 @@ award_emoji:
 - user
 priorities:
 - label
+prometheus_metrics:
+- project
 timelogs:
 - issue
 - merge_request

spec/lib/gitlab/import_export/safe_model_attributes.yml

@@ -555,6 +555,19 @@ ProjectCustomAttribute:
 - project_id
 - key
 - value
+PrometheusMetric:
+- id
+- created_at
+- updated_at
+- project_id
+- y_label
+- unit
+- legend
+- title
+- query
+- group
+- common
+- identifier
 Badge:
 - id
 - link_url

spec/lib/gitlab/prometheus/additional_metrics_parser_spec.rb

@@ -6,7 +6,7 @@ describe Gitlab::Prometheus::AdditionalMetricsParser do
   let(:parser_error_class) { Gitlab::Prometheus::ParsingError }
 
   describe '#load_groups_from_yaml' do
-    subject { described_class.load_groups_from_yaml }
+    subject { described_class.load_groups_from_yaml('dummy.yaml') }
 
     describe 'parsing sample yaml' do
       let(:sample_yaml) do

spec/lib/gitlab/prometheus/metric_group_spec.rb

+# frozen_string_literal: true
+
+require 'rails_helper'
+
+describe Gitlab::Prometheus::MetricGroup do
+  describe '.common_metrics' do
+    let!(:project_metric) { create(:prometheus_metric) }
+    let!(:common_metric_group_a) { create(:prometheus_metric, :common, group: :aws_elb) }
+    let!(:common_metric_group_b_q1) { create(:prometheus_metric, :common, group: :kubernetes) }
+    let!(:common_metric_group_b_q2) { create(:prometheus_metric, :common, group: :kubernetes) }
+
+    subject { described_class.common_metrics }
+
+    it 'returns exactly two groups' do
+      expect(subject.map(&:name)).to contain_exactly(
+        'Response metrics (AWS ELB)', 'System metrics (Kubernetes)')
+    end
+
+    it 'returns exactly three metric queries' do
+      expect(subject.map(&:metrics).flatten.map(&:id)).to contain_exactly(
+        common_metric_group_a.id, common_metric_group_b_q1.id,
+        common_metric_group_b_q2.id)
+    end
+  end
+
+  describe '.for_project' do
+    let!(:other_project) { create(:project) }
+    let!(:project_metric) { create(:prometheus_metric) }
+    let!(:common_metric) { create(:prometheus_metric, :common, group: :aws_elb) }
+
+    subject do
+      described_class.for_project(other_project)
+        .map(&:metrics).flatten
+        .map(&:id)
+    end
+
+    it 'returns exactly one common metric' do
+      is_expected.to contain_exactly(common_metric.id)
+    end
+  end
+end

spec/migrations/import_common_metrics_spec.rb

+# frozen_string_literal: true
+
+require 'spec_helper'
+require Rails.root.join('db', 'migrate', '20180831164910_import_common_metrics.rb')
+
+describe ImportCommonMetrics, :migration do
+  describe '#up' do
+    it "imports all prometheus metrics" do
+      expect(PrometheusMetric.common).to be_empty
+
+      migrate!
+
+      expect(PrometheusMetric.common).not_to be_empty
+    end
+  end
+end

spec/models/prometheus_metric_spec.rb

+# frozen_string_literal: true
+
+require 'spec_helper'
+
+describe PrometheusMetric do
+  subject { build(:prometheus_metric) }
+  let(:other_project) { build(:project) }
+
+  it { is_expected.to belong_to(:project) }
+  it { is_expected.to validate_presence_of(:title) }
+  it { is_expected.to validate_presence_of(:query) }
+  it { is_expected.to validate_presence_of(:group) }
+
+  describe 'common metrics' do
+    using RSpec::Parameterized::TableSyntax
+
+    where(:common, :project, :result) do
+      false | other_project | true
+      false | nil           | false
+      true  | other_project | false
+      true  | nil           | true
+    end
+
+    with_them do
+      before do
+        subject.common = common
+        subject.project = project
+      end
+
+      it { expect(subject.valid?).to eq(result) }
+    end
+  end
+
+  describe '#query_series' do
+    using RSpec::Parameterized::TableSyntax
+
+    where(:legend, :type) do
+      'Some other legend' | NilClass
+      'Status Code'       | Array
+    end
+
+    with_them do
+      before do
+        subject.legend = legend
+      end
+
+      it { expect(subject.query_series).to be_a(type) }
+    end
+  end
+
+  describe '#group_title' do
+    shared_examples 'group_title' do |group, title|
+      subject { build(:prometheus_metric, group: group).group_title }
+
+      it "returns text #{title} for group #{group}" do
+        expect(subject).to eq(title)
+      end
+    end
+
+    it_behaves_like 'group_title', :business, 'Business metrics (Custom)'
+    it_behaves_like 'group_title', :response, 'Response metrics (Custom)'
+    it_behaves_like 'group_title', :system, 'System metrics (Custom)'
+  end
+
+  describe '#to_query_metric' do
+    it 'converts to queryable metric object' do
+      expect(subject.to_query_metric).to be_instance_of(Gitlab::Prometheus::Metric)
+    end
+
+    it 'queryable metric object has title' do
+      expect(subject.to_query_metric.title).to eq(subject.title)
+    end
+
+    it 'queryable metric object has y_label' do
+      expect(subject.to_query_metric.y_label).to eq(subject.y_label)
+    end
+
+    it 'queryable metric has no required_metric' do
+      expect(subject.to_query_metric.required_metrics).to eq([])
+    end
+
+    it 'queryable metric has weight 0' do
+      expect(subject.to_query_metric.weight).to eq(0)
+    end
+
+    it 'queryable metrics has query description' do
+      queries = [
+        {
+          query_range: subject.query,
+          unit: subject.unit,
+          label: subject.legend
+        }
+      ]
+
+      expect(subject.to_query_metric.queries).to eq(queries)
+    end
+  end
+end