Cherry-pick #9657 to 6.x: Add redis key metricset (#9885)

Add redis key metricset to collect information about keys
like length, type or TTL.

(cherry picked from commit 5dbe803)

Add a dashboard with details for Redis keys using the new key metricset.
Visualizations included show lists length, what is useful when redis is used
for queuing, and average strings sizes and TTLs, what is useful when Redis
is used as a cache.

(cherry picked from commit ac04c32)

CHANGELOG.next.asciidoc

@@ -26,7 +26,6 @@ https://github.com/elastic/beats/compare/v6.6.0...6.x[Check the HEAD diff]
 
 *Metricbeat*
 
-- Fix issue preventing diskio metrics collection for idle disks. {issue}9124[9124] {pull}9125[9125]
 - Allow beats to blacklist certain part of the configuration while using Central Management. {pull}9099[9099]
 
 *Packetbeat*
@@ -56,6 +55,7 @@ https://github.com/elastic/beats/compare/v6.6.0...6.x[Check the HEAD diff]
 *Auditbeat*
 
 *Filebeat*
+
 - Correctly parse `December` or `Dec` in the Syslog input. {pull}9349[9349]
 - Fix improperly set config for CRI Flag in Docker Input {pull}8899[8899]
 - Just enabling the `elasticsearch` fileset and starting Filebeat no longer causes an error. {pull}8891[8891]
@@ -76,13 +76,14 @@ https://github.com/elastic/beats/compare/v6.6.0...6.x[Check the HEAD diff]
 - Add missing journalbeat non breaking fixes. {pull}9106[9106]
 
 *Metricbeat*
-- Fix panics in vsphere module when certain values where not returned by the API. {pull}9784[9784]
 
+- Fix panics in vsphere module when certain values where not returned by the API. {pull}9784[9784]
 - Add missing namespace field in http server metricset {pull}7890[7890]
 - Fix race condition when enriching events with kubernetes metadata. {issue}9055[9055] {issue}9067[9067]
 - Fix panic on docker healthcheck collection on dockers without healthchecks. {pull}9171[9171]
 - Fix issue with not collecting Elasticsearch cross-cluster replication stats correctly. {pull}9179[9179]
 - The `node.name` field in the `elasticsearch/node` metricset now correctly reports the Elasticsarch node name. Previously this field was incorrectly reporting the node ID instead. {pull}9209[9209]
+- Fix issue preventing diskio metrics collection for idle disks. {issue}9124[9124] {pull}9125[9125]
 
 *Packetbeat*
 
@@ -95,8 +96,8 @@ https://github.com/elastic/beats/compare/v6.6.0...6.x[Check the HEAD diff]
 ==== Added
 
 *Affecting all Beats*
-- Unify dashboard exporter tools. {pull}9097[9097]
 
+- Unify dashboard exporter tools. {pull}9097[9097]
 - Dissect will now flag event on parsing error. {pull}8751[8751]
 - Added the `redirect_stderr` option that allows panics to be logged to log files. {pull}8430[8430]
 - Add cache.ttl to add_host_metadata. {pull}9359[9359]
@@ -109,6 +110,7 @@ https://github.com/elastic/beats/compare/v6.6.0...6.x[Check the HEAD diff]
 - Add system module. {pull}9546[9546]
 
 *Filebeat*
+
 - Added `detect_null_bytes` selector to detect null bytes from a io.reader. {pull}9210[9210]
 - Added `syslog_host` variable to HAProxy module to allow syslog listener to bind to configured host. {pull}9366[9366]
 - Added support on Traefik for Common Log Format and Combined Log Format mixed which is the default Traefik format {issue}8015[8015] {issue}6111[6111] {pull}8768[8768].
@@ -142,6 +144,7 @@ https://github.com/elastic/beats/compare/v6.6.0...6.x[Check the HEAD diff]
 - Add field `event.dataset` which is `{module}.{metricset).
 - Add more TCP statuses to `socket_summary` metricset. {pull}9430[9430]
 - Remove experimental tag from ceph metricsets. {pull}9708[9708]
+- Add `key` metricset to the Redis module. {issue}9582[9582] {pull}9657[9657]
 
 *Packetbeat*
 

metricbeat/docs/fields.asciidoc

@@ -17693,6 +17693,63 @@ type: long
 Count of slow operations
 
 
+--
+
+[float]
+== key fields
+
+`key` contains information about keys.
+
+
+
+*`redis.key.name`*::
++
+--
+type: keyword
+
+Key name.
+
+
+--
+
+*`redis.key.id`*::
++
+--
+type: keyword
+
+Unique id for this key (With the form <keyspace>:<name>).
+
+
+--
+
+*`redis.key.type`*::
++
+--
+type: keyword
+
+Key type as shown by `TYPE` command.
+
+
+--
+
+*`redis.key.length`*::
++
+--
+type: long
+
+Length of the key (Number of elements for lists, length for strings, cardinality for sets).
+
+
+--
+
+*`redis.key.expire.ttl`*::
++
+--
+type: long
+
+Seconds to expire.
+
+
 --
 
 [float]

metricbeat/docs/modules/redis.asciidoc

@@ -81,9 +81,13 @@ The following metricsets are available:
 
 * <<metricbeat-metricset-redis-info,info>>
 
+* <<metricbeat-metricset-redis-key,key>>
+
 * <<metricbeat-metricset-redis-keyspace,keyspace>>
 
 include::redis/info.asciidoc[]
 
+include::redis/key.asciidoc[]
+
 include::redis/keyspace.asciidoc[]
 

metricbeat/docs/modules/redis/key.asciidoc

@@ -0,0 +1,21 @@
+////
+This file is generated! See scripts/docs_collector.py
+////
+
+[[metricbeat-metricset-redis-key]]
+=== Redis key metricset
+
+include::../../../module/redis/key/_meta/docs.asciidoc[]
+
+
+==== Fields
+
+For a description of each field in the metricset, see the
+<<exported-fields-redis,exported fields>> section.
+
+Here is an example document generated by this metricset:
+
+[source,json]
+----
+include::../../../module/redis/key/_meta/data.json[]
+----

metricbeat/docs/modules_list.asciidoc

@@ -119,7 +119,8 @@ This file is generated! See scripts/docs_collector.py
 |<<metricbeat-metricset-rabbitmq-node,node>> beta[]  
 |<<metricbeat-metricset-rabbitmq-queue,queue>> beta[]  
 |<<metricbeat-module-redis,Redis>>     |image:./images/icon-yes.png[Prebuilt dashboards are available]    |  
-.2+| .2+|  |<<metricbeat-metricset-redis-info,info>>   
+.3+| .3+|  |<<metricbeat-metricset-redis-info,info>>   
+|<<metricbeat-metricset-redis-key,key>>   
 |<<metricbeat-metricset-redis-keyspace,keyspace>>   
 |<<metricbeat-module-system,System>>     |image:./images/icon-yes.png[Prebuilt dashboards are available]    |  
 .14+| .14+|  |<<metricbeat-metricset-system-core,core>>