Manage Cloudkitty Elasticsearch index with template

ansible/group_vars/all.yml

@@ -953,13 +953,16 @@ cinder_backup_mount_options_nfs: ""
 #######################
 # Valid option is gnocchi
 cloudkitty_collector_backend: "gnocchi"
-# Valid options are 'sqlalchemy' or 'influxdb'. The default value is
+# Valid options are 'sqlalchemy', 'elasticsearch' or 'influxdb'. The default value is
 # 'influxdb', which matches the default in Cloudkitty since the Stein release.
 # When the backend is "influxdb", we also enable Influxdb.
-# Also, when using 'influxdb' as the backend, we trigger the configuration/use
+# Also, when using 'influxdb' or 'elasticsearch' as the backend, we trigger the configuration/use
 # of Cloudkitty storage backend version 2.
 cloudkitty_storage_backend: "influxdb"
 
+# Elasticsearch index alias name when cloudkitty_storage_backend is 'elasticsearch'.
+cloudkitty_elasticsearch_alias_name: "cloudkitty"
+
 #######################
 # Designate options
 #######################

ansible/roles/cloudkitty/defaults/main.yml

@@ -152,8 +152,21 @@ cloudkitty_storage_backend: "influxdb"
 
 cloudkitty_influxdb_name: "cloudkitty"
 
-# Set the elasticsearch index name.
-cloudkitty_elasticsearch_index_name: "cloudkitty"
+# Set the elasticsearch index alias name.
+cloudkitty_elasticsearch_alias_name: "cloudkitty"
+
+# Set the elasticsearch index name, will be URI encoded. eg, <cloudkitty-{now/d}-000001>.
+cloudkitty_elasticsearch_index_name: "<{{ cloudkitty_elasticsearch_alias_name }}-{now/d}-000001>"
+
+# Set the name of the elasticsearch component template to be created.
+cloudkitty_elasticsearch_component_template_name: "{{ cloudkitty_elasticsearch_alias_name }}_settings"
+
+# Set the elasticsearch component templates to include in the cloudkitty provided template.
+cloudkitty_elasticsearch_component_templates:
+  - "{{ cloudkitty_elasticsearch_component_template_name }}"
+
+# Set the number of replicas for cloudkitty indices.
+cloudkitty_elasticsearch_index_replicas: "{{ groups['elasticsearch'] | length -1 }}"
 
 # Set the elasticsearch host URL.
 cloudkitty_elasticsearch_url: "{{ internal_protocol }}://{{ elasticsearch_address }}:{{ elasticsearch_port }}"

ansible/roles/cloudkitty/tasks/bootstrap.yml

@@ -45,34 +45,96 @@
   delegate_to: "{{ groups['cloudkitty-api'][0] }}"
   when: cloudkitty_storage_backend == 'influxdb'
 
-- name: Checking if Cloudkitty elasticsearch index exists
+- name: Creating Cloudkitty elasticsearch index settings component template
   become: true
   kolla_toolbox:
     module_name: uri
     module_args:
-      url: "{{ cloudkitty_elasticsearch_url }}/{{ cloudkitty_elasticsearch_index_name }}"
+      url: "{{ cloudkitty_elasticsearch_url }}/_component_template/{{ cloudkitty_elasticsearch_component_template_name }}"
+      method: PUT
+      status_code: 200
+      return_content: yes
+      body: "{{ lookup('template', item) | to_json }}"
+      body_format: json
+  run_once: True
+  delegate_to: "{{ groups['cloudkitty-api'][0] }}"
+  when: cloudkitty_storage_backend == 'elasticsearch'
+  with_first_found:
+    - "{{ node_custom_config }}/cloudkitty/cloudkitty_elasticsearch_component_template.json"
+    - "templates/cloudkitty_elasticsearch_component_template.json.j2"
+
+- name: Checking if Cloudkitty elasticsearch index alias name exists
+  become: true
+  kolla_toolbox:
+    module_name: uri
+    module_args:
+      url: "{{ cloudkitty_elasticsearch_url }}/{{ cloudkitty_elasticsearch_alias_name }}"
       status_code: 200, 404
   run_once: true
   delegate_to: "{{ groups['cloudkitty-api'][0] }}"
   register: cloudkitty_index
   when: cloudkitty_storage_backend == 'elasticsearch'
 
-- name: Creating Cloudkitty elasticsearch index
+- name: Checking if Cloudkitty elasticsearch index alias name is an alias
   become: true
   kolla_toolbox:
     module_name: uri
     module_args:
-      url: "{{ cloudkitty_elasticsearch_url }}/{{ cloudkitty_elasticsearch_index_name }}"
-      method: PUT
+      url: "{{ cloudkitty_elasticsearch_url }}/_alias/{{ cloudkitty_elasticsearch_alias_name }}"
+      method: HEAD
+      status_code: 200, 404
+  run_once: true
+  delegate_to: "{{ groups['cloudkitty-api'][0] }}"
+  register: cloudkitty_index_alias
+  when:
+    - cloudkitty_index is not skipped
+    - cloudkitty_index.get('status') == 200
+
+- name: Reindexing old Cloudkitty index to new index name
+  become: true
+  kolla_toolbox:
+    module_name: uri
+    module_args:
+      url: "{{ cloudkitty_elasticsearch_url }}/_reindex"
+      method: POST
+      status_code: 200
+      body: "{{ lookup('template', 'cloudkitty_elasticsearch_reindex.json.j2') | to_json }}"
+      body_format: json
+  run_once: true
+  delegate_to: "{{ groups['cloudkitty-api'][0] }}"
+  register: cloudkitty_reindex
+  when:
+    - cloudkitty_index_alias is not skipped
+    - cloudkitty_index_alias.get('status') == 404
+
+- name: Deleting old Cloudkitty index
+  become: true
+  kolla_toolbox:
+    module_name: uri
+    module_args:
+      url: "{{ cloudkitty_elasticsearch_url }}/{{ cloudkitty_elasticsearch_alias_name }}"
+      method: DELETE
+      status_code: 200
+  run_once: true
+  delegate_to: "{{ groups['cloudkitty-api'][0] }}"
+  register: cloudkitty_delete
+  when:
+    - cloudkitty_reindex is not skipped
+
+- name: Adding alias to new Cloudkitty elasticsearch index
+  become: true
+  kolla_toolbox:
+    module_name: uri
+    module_args:
+      url: "{{ cloudkitty_elasticsearch_url }}/{{ cloudkitty_elasticsearch_index_name | urlencode | regex_replace('/','%2F') }}/_alias/{{ cloudkitty_elasticsearch_alias_name }}"
+      method: POST
       status_code: 200
       return_content: yes
-      body: |
-        {}
+      body: '{ "is_write_index": true }'
       body_format: json
   run_once: True
   delegate_to: "{{ groups['cloudkitty-api'][0] }}"
   when:
-    - cloudkitty_storage_backend == 'elasticsearch'
-    - cloudkitty_index.get('status') != 200
+    - cloudkitty_delete is not skipped
 
 - import_tasks: bootstrap_service.yml

ansible/roles/cloudkitty/templates/cloudkitty.conf.j2

@@ -126,7 +126,8 @@ cafile = {{ cloudkitty_influxdb_cafile }}
 {% if cloudkitty_storage_backend == 'elasticsearch' %}
 [storage_elasticsearch]
 host = {{ cloudkitty_elasticsearch_url }}
-index_name = {{ cloudkitty_elasticsearch_index_name }}
+index_name = {{ cloudkitty_elasticsearch_alias_name }}
+component_templates = {{ cloudkitty_elasticsearch_component_templates | join(',') }}
 insecure = {{ cloudkitty_elasticsearch_insecure_connections }}
 
 {% if cloudkitty_elasticsearch_cafile is defined %}

ansible/roles/cloudkitty/templates/cloudkitty_elasticsearch_component_template.json.j2

@@ -0,0 +1,9 @@
+{
+  "template": {
+    "settings": {
+      "index" : {
+        "number_of_replicas": {{ cloudkitty_elasticsearch_index_replicas }}
+      }
+    }
+  }
+}

ansible/roles/cloudkitty/templates/cloudkitty_elasticsearch_reindex.json.j2

@@ -0,0 +1,8 @@
+{
+  "source": {
+    "index": "{{ cloudkitty_elasticsearch_alias_name }}"
+  },
+  "dest": {
+    "index": "{{ cloudkitty_elasticsearch_index_name }}"
+  }
+}

ansible/roles/elasticsearch/defaults/main.yml

@@ -62,9 +62,15 @@ elasticsearch_curator_cron_schedule: "0 {{ elasticsearch_curator_instance_id }}
 # useful way of checking that Curator actions are working as expected.
 elasticsearch_curator_dry_run: false
 
+# List of elasticsearch index names to be managed by Curator.
+elasticsearch_curator_index_names:
+  - "{{ kibana_log_prefix }}"
+  - "{{ (enable_monasca | bool) | ternary('monasca', '') }}"
+  - "{{ (cloudkitty_storage_backend == 'elasticsearch') | ternary(cloudkitty_elasticsearch_alias_name, '') }}"
+
 # Index prefix pattern. Any indices matching this regex will
 # be managed by Curator.
-elasticsearch_curator_index_pattern: "^{{ '(monasca|' + kibana_log_prefix + ')' if enable_monasca|bool else kibana_log_prefix }}-.*"
+elasticsearch_curator_index_pattern: "^({{ elasticsearch_curator_index_names | reject('eq', '') | unique | join('|') }})-.*"
 
 # Duration after which an index is staged for deletion. This is
 # implemented by closing the index. Whilst in this state the index
@@ -75,6 +81,24 @@ elasticsearch_curator_soft_retention_period_days: 30
 # Duration after which an index is permanently erased from the cluster.
 elasticsearch_curator_hard_retention_period_days: 60
 
+# Enable rollover of elasticsearch indices.
+elasticsearch_curator_rollover: "{{ cloudkitty_storage_backend == 'elasticsearch' }}"
+
+# Dict containing elasticsearch rollover options.
+# For help with options see curator docs:
+# https://www.elastic.co/guide/en/elasticsearch/client/curator/5.8/rollover.html
+elasticsearch_curator_rollover_options:
+  - name: "{{ cloudkitty_elasticsearch_alias_name }}"
+    conditions:
+      max_age: "{{ elasticsearch_curator_rollover_default_max_age }}"
+      max_docs: "{{ elasticsearch_curator_rollover_default_max_docs }}"
+
+# Elasticsearch index max age before rollover.
+elasticsearch_curator_rollover_default_max_age: "1d"
+
+# Elasticsearch index max number of docs before rollover.
+elasticsearch_curator_rollover_default_max_docs: "100000"
+
 ####################
 # Keystone
 ####################

ansible/roles/elasticsearch/templates/elasticsearch-curator-actions.yml.j2

@@ -31,3 +31,13 @@ actions:
       timestring: '%Y.%m.%d'
       unit: days
       unit_count: "{{ elasticsearch_curator_hard_retention_period_days }}"
+{% if elasticsearch_curator_rollover | bool %}
+{% for index in elasticsearch_curator_rollover_options %}
+  {{ 2 + loop.index }}:
+    action: rollover
+    description: >-
+      Rollover index associated with {{ index.name }} alias
+    options:
+      {{ index | to_nice_yaml | indent(6, false) }}
+{% endfor %}
+{% endif %}

releasenotes/notes/add-cloudkitty-elasticsearch-alias-support-e6cc3fe1360aa4ae.yaml

@@ -0,0 +1,13 @@
+---
+features:
+  - |
+    Improves support for elasticsearch as the cloudkitty storage backend.
+    Enforces using an elasticsearch index alias with cloudkitty.
+    Provides an elasticsearch index component template to manage replicas.
+    Adds elasicsearch curator config to prune cloudkitty indices.
+upgrade:
+  - |
+    Enforces using an elasticsearch index alias with cloudkitty.
+    If elasticsearch is used as the cloudkitty storage backend this will
+    reindex the data into a dated index and delete the old copy. The old index
+    name is used as an alias for the dated indices.