Optimize Project related count service desk enabled

What does this MR do?

Optimize query produced by https://gitlab.com/gitlab-org/gitlab/-/blob/908902d1d4f6fbcd7780150c780837fd4fa8b301/ee/lib/ee/gitlab/usage_data.rb#L103

query with batching

SELECT
    COUNT("projects"."id")
FROM
    "projects"
WHERE
    "projects"."service_desk_enabled" = TRUE
    AND "projects"."id" BETWEEN 10000 AND 11250;

explain data from database-lab

• id between 10_000 AND 11_250

Before

• https://explain.depesz.com/s/mRKD

Observations

1. Filter being applied on service_desk_enabled

Index Cond: ((projects.id >= 10000) AND (projects.id <= 11250))
Filter: projects.service_desk_enabled

• combat this with adding index

CREATE INDEX CONCURRENTLY index_projects_on_id_sd ON public.projects USING btree (id)
WHERE
    service_desk_enabled = TRUE;

• explain now shows index only condition hit for that part
  • https://explain.depesz.com/s/7gKk

2. MAX/MIN calculations

• Query

SELECT
    MAX("projects"."id")
FROM
    "projects"
WHERE
    "projects"."service_desk_enabled" = TRUE;

• Explain results(no index, within acceptable thresholds - 7ms) - https://explain.depesz.com/s/2wtY
  • After adding indexes for count itself - https://explain.depesz.com/s/O9dc

Plan

• Add these indexes

CREATE INDEX CONCURRENTLY index_projects_on_id_sd ON public.projects USING btree (id)
WHERE
    service_desk_enabled = TRUE;

After only index conditions are hit

• https://explain.depesz.com/s/7gKk

Timing

After the index for batch counting takes 2.2 seconds super pessimistic using database-lab

• 5.5 million users,
• with 1_250 batch sizes
• 5.5M/1_250 = 4_400 loops
• Time: < 0.5ms ( cold cache )

Migration output

12:34 $ be rails db:migrate:up VERSION=20200312163407
== 20200312163407 AddIndexOnIdAndServiceDeskEnabledToProjects: migrating ======
-- transaction_open?()
   -> 0.0000s
-- index_exists?(:projects, :id, {:where=>"service_desk_enabled = true", :name=>"index_projects_on_id_service_desk_enabled", :algorithm=>:concurrently})
   -> 0.0114s
-- execute("SET statement_timeout TO 0")
   -> 0.0004s
-- add_index(:projects, :id, {:where=>"service_desk_enabled = true", :name=>"index_projects_on_id_service_desk_enabled", :algorithm=>:concurrently})
   -> 0.0053s
-- execute("RESET ALL")
   -> 0.0004s
== 20200312163407 AddIndexOnIdAndServiceDeskEnabledToProjects: migrated (0.0176s)

✔ ~/projects/gdk/gitlab [208887-optimize-project-counters-service-desk ↑·2|✚ 1…1⚑ 1]
12:47 $ be rails db:migrate:down VERSION=20200312163407
== 20200312163407 AddIndexOnIdAndServiceDeskEnabledToProjects: reverting ======
-- transaction_open?()
   -> 0.0000s
-- indexes(:projects)
   -> 0.0114s
-- execute("SET statement_timeout TO 0")
   -> 0.0004s
-- remove_index(:projects, {:algorithm=>:concurrently, :name=>"index_projects_on_id_service_desk_enabled"})
   -> 0.0025s
-- execute("RESET ALL")
   -> 0.0004s
== 20200312163407 AddIndexOnIdAndServiceDeskEnabledToProjects: reverted (0.0148s)

Does this MR meet the acceptance criteria?

Conformity

• Changelog entry
• Documentation (if required)
• Code review guidelines
• Merge request performance guidelines
• Style guides
• Database guides
• Separation of EE specific content

Availability and Testing

• Review and add/update tests for this feature/bug. Consider all test levels. See the Test Planning Process.
• Tested in all supported browsers
• Informed Infrastructure department of a default or new setting change, if applicable per definition of done

Security

If this MR contains changes to processing or storing of credentials or tokens, authorization and authentication methods and other items described in the security review guidelines:

• Label as security and @ mention @gitlab-com/gl-security/appsec
• The MR includes necessary changes to maintain consistency between UI, API, email, or other methods
• Security reports checked/validated by a reviewer from the AppSec team

Related to #208887 (closed)