Update documentation and add new Rake task:

- Added yum command for libicu-devel install
- Added `git clone` and cd commands to install instructions
- Added note about system requirements
- Added optional step to delete the index for disabling ElasticSearch
- Removed step about creating empty index, as wrapper script does this
- Added warning about async indexing and sidekiq
- Added section about rake tasks
  - Defined rake tasks and linked to corresponding rake file
- Added periods to the end of rake task descriptions
- Added Scopes section
- Added Environment Variables Section
  - Added subsection about batching
  - Added subsection about indexing one project
- Added more to troubleshooting not finding data
- Added more to troubleshooting concerning time it takes to index
- Added troubleshooting section verifying ElasticSearch is being used
- Fixed typo in indexing small instances section
  - use the index the data ~>  index the data
- Added hyperlink to the Enabling ElasticSearch section
- Removed unneeded comma from line 33
- Put emphasis on MUST on line 173
- Changed some wording on line 187
  - time drop -> decreae in indexing time
- Fixed spelling and grammar mistakes
- Added example of using BATCH
- Fixed typo on line 67: TO ~> To
- Removed `the` on line 442
- Wrapped scope names in backtickets on lines 428-436
- Changed `user` to `GitLab instance` on line 402
- Specfied default 300 projects instead of just 300 on line 384 and 392
- Removed line numbers from links, lines 332-376
- Changed * to - for listing style, lines 332-376
- Changed * to - for listing style, lines 397-400
- Added backticks to variable names, lines 384-387
- Removed a `#` from Batching for spacing niceness, line 389
- Removed a `#` from Indexing a specific... for spacing niceness, 412
- Added periods for full stops on env variable table, lines 384-387
- Changed pipe symbol to slash for table formatting on L385
- Changed all instances of ElasticSearch to Elasticsearch
- Changed Project.all to ActiveRecord query
- Used count on ActiveRecord query to determine if all were indexed
- Used find_each to process in batches for efficiency
- Changed ActiveRecord query to make it cleaner
- Removed spec test for new rake task as it was unneeded
- Remade projects_not_indexed task to show 1st 500 nonindexed projects
  - Added ability to display all using env variable SHOW_ALL
- Put a newline before `if not_indexed`
- Put a newline between `end` and `arr`
- Put a newline between `end` and `puts`
- Fixed spacing issues
- Used return of conditional for vairable assignment on L224
- Added README.md to high_availability link

ee/changelogs/unreleased/10109-rake-task-for-elasticsearch-to-determine-which-projects-are-not-indexed.yml

+---
+title: Added gitlab:elastic:projects_not_indexed rake task
+merge_request: 9854
+author: Jason Colyer
+type: added

ee/lib/tasks/gitlab/elastic.rake

@@ -179,6 +179,17 @@ namespace :gitlab do
       puts "Done".color(:green)
     end
 
+    desc "GitLab | Elasticsearch | Display which projects are not indexed"
+    task projects_not_indexed: :environment do
+      not_indexed = Project.where.not(id: IndexStatus.select(:project_id).distinct)
+
+      if not_indexed.count.zero?
+        puts 'All projects are currently indexed'.color(:green)
+      else
+        display_unindexed(not_indexed)
+      end
+    end
+
     def batch_size
       ENV.fetch('BATCH', 300).to_i
     end
@@ -208,5 +219,19 @@ namespace :gitlab do
 
       projects
     end
+
+    def display_unindexed(projects)
+      arr = if projects.count < 500 || ENV['SHOW_ALL']
+              projects
+            else
+              projects[1..500]
+            end
+
+      arr.each do |p|
+        puts "Project '#{p.full_path}' (ID: #{p.id}) isn't indexed.".color(:red)
+      end
+
+      puts "#{arr.count} out of #{projects.count} non-indexed projects shown."
+    end
   end
 end