Merge branch '12375-elastic-bulk-config' into 'master'

Make elasticsearch bulk parameters configurable

See merge request gitlab-org/gitlab!24688

GITLAB_ELASTICSEARCH_INDEXER_VERSION

-2.0.0
+2.1.0

db/migrate/20200207132752_add_es_bulk_config.rb

+# frozen_string_literal: true
+
+class AddEsBulkConfig < ActiveRecord::Migration[6.0]
+  # Set this constant to true if this migration requires downtime.
+  DOWNTIME = false
+
+  def change
+    add_column :application_settings, :elasticsearch_max_bulk_size_mb, :smallint, null: false, default: 10
+    add_column :application_settings, :elasticsearch_max_bulk_concurrency, :smallint, null: false, default: 10
+  end
+end

db/schema.rb

@@ -344,6 +344,8 @@ ActiveRecord::Schema.define(version: 2020_02_11_152410) do
     t.boolean "updating_name_disabled_for_users", default: false, null: false
     t.integer "instance_administrators_group_id"
     t.integer "elasticsearch_indexed_field_length_limit", default: 0, null: false
+    t.integer "elasticsearch_max_bulk_size_mb", limit: 2, default: 10, null: false
+    t.integer "elasticsearch_max_bulk_concurrency", limit: 2, default: 10, null: false
     t.index ["custom_project_templates_group_id"], name: "index_application_settings_on_custom_project_templates_group_id"
     t.index ["file_template_project_id"], name: "index_application_settings_on_file_template_project_id"
     t.index ["instance_administration_project_id"], name: "index_applicationsettings_on_instance_administration_project_id"

doc/integration/elasticsearch.md

@@ -151,6 +151,8 @@ The following Elasticsearch settings are available:
 | `AWS Access Key`                                      | The AWS access key. |
 | `AWS Secret Access Key`                               | The AWS secret access key. |
 | `Maximum field length`                                | See [the explanation in instance limits.](../administration/instance_limits.md#maximum-field-length). |
+| `Maximum bulk request size (MiB)` | Repository indexing uses the Elasticsearch bulk request API. This setting determines the maximum size of an individual bulk request during these operations. |
+| `Bulk request concurrency` | Each repository indexing operation may submit bulk requests in parallel. This increases indexing performance, but fills the Elasticsearch bulk requests queue faster. |
 
 ### Limiting namespaces and projects
 

ee/app/helpers/ee/application_settings_helper.rb

@@ -26,6 +26,8 @@ module EE
         :elasticsearch_aws_region,
         :elasticsearch_aws_secret_access_key,
         :elasticsearch_indexing,
+        :elasticsearch_max_bulk_concurrency,
+        :elasticsearch_max_bulk_size_mb,
         :elasticsearch_replicas,
         :elasticsearch_indexed_field_length_limit,
         :elasticsearch_search,

ee/app/models/ee/application_setting.rb

@@ -48,6 +48,14 @@ module EE
                 presence: true,
                 numericality: { only_integer: true, greater_than: 0 }
 
+      validates :elasticsearch_max_bulk_size_mb,
+                presence: true,
+                numericality: { only_integer: true, greater_than: 0 }
+
+      validates :elasticsearch_max_bulk_concurrency,
+                presence: true,
+                numericality: { only_integer: true, greater_than: 0 }
+
       validates :elasticsearch_url,
                 presence: { message: "can't be blank when indexing is enabled" },
                 if: ->(setting) { setting.elasticsearch_indexing? }
@@ -90,6 +98,8 @@ module EE
           elasticsearch_replicas: 1,
           elasticsearch_shards: 5,
           elasticsearch_indexed_field_length_limit: 0,
+          elasticsearch_max_bulk_size_bytes: 10.megabytes,
+          elasticsearch_max_bulk_concurrency: 10,
           elasticsearch_url: ENV['ELASTIC_URL'] || 'http://localhost:9200',
           email_additional_text: nil,
           lock_memberships_to_ldap: false,
@@ -209,7 +219,9 @@ module EE
         aws:                   elasticsearch_aws,
         aws_access_key:        elasticsearch_aws_access_key,
         aws_secret_access_key: elasticsearch_aws_secret_access_key,
-        aws_region:            elasticsearch_aws_region
+        aws_region:            elasticsearch_aws_region,
+        max_bulk_size_bytes:   elasticsearch_max_bulk_size_mb.megabytes,
+        max_bulk_concurrency:  elasticsearch_max_bulk_concurrency
       }
     end
 

ee/app/views/admin/application_settings/_elasticsearch_form.html.haml

@@ -63,6 +63,20 @@
             .form-text.text-muted
               = _('If any indexed field exceeds this limit it will be truncated to this number of characters and the rest will not be indexed or searchable. This does not apply to repository and wiki indexing. Setting this to 0 means it is unlimited.')
 
+          .form-group
+            = f.label :elasticsearch_max_bulk_size_mb, _('Maximum bulk request size (MiB)'), class: 'label-bold'
+            = f.number_field :elasticsearch_max_bulk_size_mb, value: @application_setting.elasticsearch_max_bulk_size_mb, class: 'form-control'
+            .form-text.text-muted
+              = _('Maximum size of Elasticsearch bulk indexing requests.')
+              = _('This only applies to repository indexing operations.')
+
+          .form-group
+            = f.label :elasticsearch_max_bulk_concurrency, _('Bulk request concurrency'), class: 'label-bold'
+            = f.number_field :elasticsearch_max_bulk_concurrency, value: @application_setting.elasticsearch_max_bulk_concurrency, class: 'form-control'
+            .form-text.text-muted
+              = _('Maximum concurrency of Elasticsearch bulk requests per indexing operation.')
+              = _('This only applies to repository indexing operations.')
+
         .sub-section
           %h4= _('Elasticsearch indexing restrictions')
           .form-group

ee/changelogs/unreleased/12375-elastic-bulk-config.yml

+---
+title: Make elasticsearch bulk parameters configurable
+merge_request: 24688
+author:
+type: added

ee/spec/features/admin/admin_settings_spec.rb

@@ -69,6 +69,8 @@ describe 'Admin updates EE-only settings' do
         fill_in 'Number of Elasticsearch shards', with: '120'
         fill_in 'Number of Elasticsearch replicas', with: '2'
         fill_in 'Maximum field length', with: '100000'
+        fill_in 'Maximum bulk request size (MiB)', with: '17'
+        fill_in 'Bulk request concurrency', with: '23'
 
         click_button 'Save changes'
       end
@@ -79,6 +81,8 @@ describe 'Admin updates EE-only settings' do
         expect(current_settings.elasticsearch_shards).to eq(120)
         expect(current_settings.elasticsearch_replicas).to eq(2)
         expect(current_settings.elasticsearch_indexed_field_length_limit).to eq(100000)
+        expect(current_settings.elasticsearch_max_bulk_size_mb).to eq(17)
+        expect(current_settings.elasticsearch_max_bulk_concurrency).to eq(23)
         expect(page).to have_content 'Application settings saved successfully'
       end
     end

ee/spec/models/application_setting_spec.rb

@@ -47,6 +47,18 @@ describe ApplicationSetting do
     it { is_expected.not_to allow_value(1.1).for(:elasticsearch_indexed_field_length_limit) }
     it { is_expected.not_to allow_value(-1).for(:elasticsearch_indexed_field_length_limit) }
 
+    it { is_expected.to allow_value(25).for(:elasticsearch_max_bulk_size_mb) }
+    it { is_expected.not_to allow_value(nil).for(:elasticsearch_max_bulk_size_mb) }
+    it { is_expected.not_to allow_value(0).for(:elasticsearch_max_bulk_size_mb) }
+    it { is_expected.not_to allow_value(1.1).for(:elasticsearch_max_bulk_size_mb) }
+    it { is_expected.not_to allow_value(-1).for(:elasticsearch_max_bulk_size_mb) }
+
+    it { is_expected.to allow_value(2).for(:elasticsearch_max_bulk_concurrency) }
+    it { is_expected.not_to allow_value(nil).for(:elasticsearch_max_bulk_concurrency) }
+    it { is_expected.not_to allow_value(0).for(:elasticsearch_max_bulk_concurrency) }
+    it { is_expected.not_to allow_value(1.1).for(:elasticsearch_max_bulk_concurrency) }
+    it { is_expected.not_to allow_value(-1).for(:elasticsearch_max_bulk_concurrency) }
+
     it { is_expected.to allow_value(nil).for(:required_instance_ci_template) }
     it { is_expected.not_to allow_value("").for(:required_instance_ci_template) }
     it { is_expected.not_to allow_value("  ").for(:required_instance_ci_template) }
@@ -208,7 +220,9 @@ describe ApplicationSetting do
         elasticsearch_aws: false,
         elasticsearch_aws_region:     'test-region',
         elasticsearch_aws_access_key: 'test-access-key',
-        elasticsearch_aws_secret_access_key: 'test-secret-access-key'
+        elasticsearch_aws_secret_access_key: 'test-secret-access-key',
+        elasticsearch_max_bulk_size_mb: 67,
+        elasticsearch_max_bulk_concurrency: 8
       )
 
       expect(setting.elasticsearch_config).to eq(
@@ -216,7 +230,9 @@ describe ApplicationSetting do
         aws: false,
         aws_region:     'test-region',
         aws_access_key: 'test-access-key',
-        aws_secret_access_key: 'test-secret-access-key'
+        aws_secret_access_key: 'test-secret-access-key',
+        max_bulk_size_bytes: 67.megabytes,
+        max_bulk_concurrency: 8
       )
     end
 

locale/gitlab.pot

@@ -2999,6 +2999,9 @@ msgstr ""
 msgid "Built-in"
 msgstr ""
 
+msgid "Bulk request concurrency"
+msgstr ""
+
 msgid "Burndown chart"
 msgstr ""
 
@@ -11728,9 +11731,15 @@ msgstr ""
 msgid "Maximum attachment size (MB)"
 msgstr ""
 
+msgid "Maximum bulk request size (MiB)"
+msgstr ""
+
 msgid "Maximum capacity"
 msgstr ""
 
+msgid "Maximum concurrency of Elasticsearch bulk requests per indexing operation."
+msgstr ""
+
 msgid "Maximum delay (Minutes)"
 msgstr ""
 
@@ -11773,6 +11782,9 @@ msgstr ""
 msgid "Maximum size limit for each repository."
 msgstr ""
 
+msgid "Maximum size of Elasticsearch bulk indexing requests."
+msgstr ""
+
 msgid "Maximum size of individual attachments in comments."
 msgstr ""
 
@@ -19578,6 +19590,9 @@ msgstr ""
 msgid "This namespace has already been taken! Please choose another one."
 msgstr ""
 
+msgid "This only applies to repository indexing operations."
+msgstr ""
+
 msgid "This option is only available on GitLab.com"
 msgstr ""