Squashed commit of the following:

commit 22e1cb8f4b98f71d21026f69aa3e68d79946d6ae
Merge: 766a42a9639 e966c6aea25
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Fri Feb 23 21:30:46 2018 +0100

    Merge remote-tracking branch 'upstream/master' into 38783-add-cluster-metrics.yml

    # Conflicts:
    #	app/controllers/projects/prometheus/metrics_controller.rb
    #	app/controllers/projects/prometheus_controller.rb
    #	app/models/project_services/prometheus_service.rb
    #	lib/gitlab/prometheus/queries/query_additional_metrics.rb
    #	spec/controllers/projects/prometheus/metrics_controller_spec.rb
    #	spec/models/project_services/prometheus_service_spec.rb

commit 766a42a96393f502d439c1f0beb0b6cfb2c228d4
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Fri Feb 23 17:58:45 2018 +0100

    Move prometheus adapter to app/models/concerns

commit ca84eed49811cf3064a2e5ea611af1c947c590d7
Merge: 66702099586 cb504cedc2a
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Fri Feb 23 15:36:03 2018 +0100

    Merge remote-tracking branch 'upstream/backport_custom_metrics_ce_components' into 38783-add-cluster-metrics.yml

    + fix failing tests

    # Conflicts:
    #	app/controllers/projects/prometheus_controller.rb
    #	app/models/project_services/prometheus_service.rb
    #	lib/gitlab/prometheus/queries/query_additional_metrics.rb
    #	spec/models/project_services/prometheus_service_spec.rb

commit cb504cedc2a6e353ffb56833334681e3da09fc14
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Fri Feb 23 14:46:15 2018 +0100

    Fix prometheus_service found by find_or_initialize_service

commit 928b84c72c2c4c46e1785b9a943c1822a137de16
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 23:15:11 2018 +0100

    additional metrics and backported tests

commit 41291383b4b6976af94eaf9e1a7b2e2a172310e7
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 21:57:45 2018 +0100

    Introduce Metrics controller and retire prometheus controller

commit 6e7492e4c7ffa9d8621f09198071bc14be875976
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 21:46:37 2018 +0100

    Backport PrometheusClient::Error and all->common_metrics rename

commit 66702099586c864a7f78970a0ee0dd9f3c2beeef
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 20:23:03 2018 +0100

    fix failing tests

commit ce921ea723cde61b2441ba98c9aca97606c719c8
Author: Mike Greiling <mike@pixelcog.com>
Date:   Wed Feb 21 23:47:30 2018 -0600

    add labels to cluster metrics

commit 993830c6892e4fbf53d3f72da00002d642c7e9b2
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 01:20:49 2018 +0100

    Fix formatting probelms and few small tests

commit 66ee65d8e3f747c90d986bc3056178422156bc8e
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 01:08:03 2018 +0100

    stop using in deployment tests environment.id

commit 62c91978d15f0369988521363dae24bd7510d68d
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 01:06:33 2018 +0100

    fix prometheus_controller and adapter tests

commit 977b1d34c1d03c7233582e8328f85caf634895ed
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 22 00:19:39 2018 +0100

    finish up active? -> can_query? rename

commit e614f7daee58a9758d83ba3efe1649c8b80bc1e2
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Wed Feb 21 23:51:33 2018 +0100

    deployment prometheus adapter tests fix

commit ebd726c114a6026fef0adf3eba6ee1972530148a
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Wed Feb 21 21:40:40 2018 +0100

    Move environment dependant tests to environment

commit 6d31311cd3729c29233283dded70e03a4a9a3c97
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Wed Feb 21 20:40:24 2018 +0100

    update monitoring service and move adding dpeloyment_time to deployment model

commit 60b6bf391ab36846dce122bc6b0c5196a186267c
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Wed Feb 21 20:15:46 2018 +0100

    adjust deployment spec and prometheus specs

commit 6681662cf1c028aff2ff94aa0501732cb7119ba1
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 22:30:20 2018 +0100

    Revert changes to reactive caching

commit e282f86c45a056889f57d3f7fd23a81c88efff6a
Merge: 5751c73df59 6844a2df873
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 22:08:17 2018 +0100

    Merge remote-tracking branch 'upstream/master' into 38783-add-cluster-metrics.yml

commit 5751c73df59d0a03840a1b4b71b0637670f971a6
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 22:07:43 2018 +0100

    rename active? to can_query? and cleanup environment prometheus router

commit 3f3c6e1d33dcd9315979daf26a95f2aab83a7de9
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 21:51:43 2018 +0100

    Fix tests, and only use prometheus service if its active

commit 6345838bac584c213b665d334252ccab202cb271
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 16:24:27 2018 +0100

    Fix typo

commit 7a585d32afe8da050b5615b1d036a550e06479f5
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 15:29:40 2018 +0100

    Cluster id is not required

commit e6af62afb11fa380f6aff1c31a81bcc9bab3b1eb
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 15:15:46 2018 +0100

    Result transformation support

commit f3b1bd7c67894f44efe33591ddb70093bd620c03
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 15:07:07 2018 +0100

    Fix rubocop warning and exten cluster query timeframe

commit be77947cea64261a4d3dead33c3c57f413a9880c
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 14:42:47 2018 +0100

    Fix additional metrics test

commit eb3922e16221abe16f59fae1c38122f227643343
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 14:34:44 2018 +0100

    rename prometheus adapter methods

commit 045476cd08b21593818b274ae8a44d19b705523f
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 13:19:36 2018 +0100

    Make prometheus adapter a module

commit f2daf050d8c689f72c4c61207930bc53c331f12e
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 20 12:00:44 2018 +0100

    refactoring wip

commit 52e4ef5587794e811dc10a0f2dca522342a865da
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Sun Feb 18 20:25:55 2018 +0100

    cleanup prometheus adapter concept

commit 3887365faab9dfcd9c00bcfc501d09ac62431a03
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Sun Feb 18 19:22:11 2018 +0100

    Refactor out deployment id, Rename PrometheusQuerier to PrometheusAdapter

commit aa2fc2df57bd72c9a5e94f66d1f1e23990be6c3f
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Fri Feb 16 23:59:54 2018 +0100

    Refactor prometheus client

commit e43c1ca9d9874d6cf1569f40fa1aca158d9d5d91
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 15 20:24:45 2018 +0100

    Use initial version of cluster_metrics.yml

commit 867821ce0b2609ebf8994220aa8e3a94d66a01e0
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 15 05:24:14 2018 +0100

    Fix querying cluster metrics

commit 1601e002a064cbb10ffe110a19433c5662858f1d
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 15 04:49:03 2018 +0100

    Queues for unicersal querier

commit 5db198fdc925c0223be24939b76da1d544dd569c
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 15 03:35:17 2018 +0100

    refactor reactive caching and prometheus querying

commit b0fc00e8c9d21e961ef44b0129103e2a62928b52
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Thu Feb 15 00:12:03 2018 +0100

    Add generic query additional metrics

commit ffe76e6a9ba196bccff22c4880e0384959ad5e48
Author: Pawel Chojnacki <pawel@chojnacki.ws>
Date:   Tue Feb 13 15:40:15 2018 +0100

    Cluster Metric yml initial

+ Remove cluster query

+ remove cluster_metrics.yml

+ Prometheus adapter tests

1 files changed, 4 insertions, 81 deletions

diff --git a/app/models/project_services/prometheus_service.rb b/app/models/project_services/prometheus_service.rb
index 58731451429..7a0c9a33e70 100644
--- a/app/models/project_services/prometheus_service.rb
+++ b/app/models/project_services/prometheus_service.rb
@@ -1,9 +1,5 @@
 class PrometheusService < MonitoringService
-  include ReactiveService
-
-  self.reactive_cache_lease_timeout = 30.seconds
-  self.reactive_cache_refresh_interval = 30.seconds
-  self.reactive_cache_lifetime = 1.minute
+  include PrometheusAdapter
 
   #  Access to prometheus is directly through the API
   prop_accessor :api_url
@@ -66,63 +62,15 @@ class PrometheusService < MonitoringService
 
   # Check we can connect to the Prometheus API
   def test(*args)
-    client.ping
+    Gitlab::PrometheusClient.new(prometheus_client).ping
 
     { success: true, result: 'Checked API endpoint' }
   rescue Gitlab::PrometheusClient::Error => err
     { success: false, result: err }
   end
 
-  def environment_metrics(environment)
-    with_reactive_cache(Gitlab::Prometheus::Queries::EnvironmentQuery.name, environment.id, &rename_field(:data, :metrics))
-  end
-
-  def deployment_metrics(deployment)
-    metrics = with_reactive_cache(Gitlab::Prometheus::Queries::DeploymentQuery.name, deployment.environment.id, deployment.id, &rename_field(:data, :metrics))
-    metrics&.merge(deployment_time: deployment.created_at.to_i) || {}
-  end
-
-  def additional_environment_metrics(environment)
-    with_reactive_cache(Gitlab::Prometheus::Queries::AdditionalMetricsEnvironmentQuery.name, environment.id, &:itself)
-  end
-
-  def additional_deployment_metrics(deployment)
-    with_reactive_cache(Gitlab::Prometheus::Queries::AdditionalMetricsDeploymentQuery.name, deployment.environment.id, deployment.id, &:itself)
-  end
-
-  def matched_metrics
-    with_reactive_cache(Gitlab::Prometheus::Queries::MatchedMetricsQuery.name, &:itself)
-  end
-
-  # Cache metrics for specific environment
-  def calculate_reactive_cache(query_class_name, *args)
-    return unless active? && project && !project.pending_delete?
-
-    environment_id = args.first
-    client = client(environment_id)
-
-    data = Kernel.const_get(query_class_name).new(client).query(*args)
-    {
-      success: true,
-      data: data,
-      last_update: Time.now.utc
-    }
-  rescue Gitlab::PrometheusClient::Error => err
-    { success: false, result: err.message }
-  end
-
-  def client(environment_id = nil)
-    if manual_configuration?
-      Gitlab::PrometheusClient.new(RestClient::Resource.new(api_url))
-    else
-      cluster = cluster_with_prometheus(environment_id)
-      raise Gitlab::PrometheusClient::Error, "couldn't find cluster with Prometheus installed" unless cluster
-
-      rest_client = client_from_cluster(cluster)
-      raise Gitlab::PrometheusClient::Error, "couldn't create proxy Prometheus client" unless rest_client
-
-      Gitlab::PrometheusClient.new(rest_client)
-    end
+  def prometheus_client
+    RestClient::Resource.new(api_url) if api_url && manual_configuration? && active?
   end
 
   def prometheus_installed?
@@ -134,31 +82,6 @@ class PrometheusService < MonitoringService
 
   private
 
-  def cluster_with_prometheus(environment_id = nil)
-    clusters = if environment_id
-                 ::Environment.find_by(id: environment_id).try do |env|
-                   # sort results by descending order based on environment_scope being longer
-                   # thus more closely matching environment slug
-                   project.clusters.enabled.for_environment(env).sort_by { |c| c.environment_scope&.length }.reverse!
-                 end
-               else
-                 project.clusters.enabled.for_all_environments
-               end
-
-    clusters&.detect { |cluster| cluster.application_prometheus&.installed? }
-  end
-
-  def client_from_cluster(cluster)
-    cluster.application_prometheus.proxy_client
-  end
-
-  def rename_field(old_field, new_field)
-    -> (metrics) do
-      metrics[new_field] = metrics.delete(old_field)
-      metrics
-    end
-  end
-
   def synchronize_service_state!
     self.active = prometheus_installed? || manual_configuration?