Edit Elastic Agent docs

x-pack/elastic-agent/docs/elastic-agent-command-line.asciidoc

@@ -0,0 +1,21 @@
+[[elastic-agent-cmd-options]]
+= Command line options
+
+experimental[]
+
+The `elastic-agent run` command provides flags that alter the behavior of an
+agent:
+
+`-path.home`::
+The home directory of the {agent}. `path.home` determines the location of the
+configuration files and data directory.
+
+`-c`::
+The configuration file to load. If not specified, {agent} uses
+`{path.home}/elastic-agent.yml`.
+
+`-path.data`::
+The data directory used by {agent} to store downloaded artifacts. Also stores
+logs for any {beats} started and managed by {agent}.
++
+If not specified, {agent} uses `{path.home}/data`.

x-pack/elastic-agent/docs/elastic-agent-configuration-example.asciidoc

@@ -0,0 +1,13 @@
+[[elastic-agent-configuration-example]]
+= Configuration example
+
+experimental[]
+
+The following example shows a full list of configuration options:
+
+//REVIEWERS: I'm not sure whether this is a true statement ^^.
+
+[source,yaml]
+----
+include::elastic-agent_configuration_example.yml[]
+----

x-pack/elastic-agent/docs/elastic-agent-configuration.asciidoc

@@ -0,0 +1,113 @@
+[[elastic-agent-configuration]]
+= Configuration settings
+
+experimental[]
+
+By default {agent} runs in standalone mode to ingest system data and send it to
+a local {es} instance running on port 9200. It uses the demo credentials of the
+`elastic` user. It's also configured to monitor all {beats} managed by the agent
+and send the {beats} logs and metrics to the same {es) instance.
+
+To alter this behavior, configure the output and other configuration settings:
+
+* <<elastic-agent-output-configuration>>
+* <<elastic-agent-monitoring-configuration>>
+* <<elastic-agent-datasource-configuration>>
+
+[float]
+[[elastic-agent-output-configuration]]
+== Output settings
+
+Specify one or more outputs. Specifying multiple outputs allows you to pair
+each data source with a different output. 
+
+IMPORTANT: {agent} currently works with the {es} output only.
+
+Example output configuration:
+
+[source,yaml]
+-------------------------------------------------------------------------------------
+outputs:
+  default:
+    type: elasticsearch
+    hosts: [127.0.0.1:9200]
+    username: elastic
+    password: changeme
+
+  monitoring:
+    type: elasticsearch
+    api_key: VuaCfGcBCdbkQm-e5aOx:ui2lp2axTNmsyakw9tvNnw
+    hosts: ["localhost:9200"]
+    ca_sha256: "7lHLiyp4J8m9kw38SJ7SURJP4bXRZv/BNxyyXkCcE/M="
+-------------------------------------------------------------------------------------
+
+This example configures two outputs: `default` and  `monitoring`.
+Notice that they use different authentication methods. The first one uses a
+username and password pair, and the second one contains an API key.
+
+[NOTE]
+==============
+A default output configuration is required.
+==============
+
+[float]
+[[elastic-agent-monitoring-configuration]]
+== {beats} monitoring settings
+
+{agent} monitors {beats} by default. To disable or change monitoring 
+settings, set options under `settings.monitoring`:
+
+[source,yaml]
+-------------------------------------------------------------------------------------
+settings.monitoring:
+  # enabled turns on monitoring of running processes
+  enabled: true
+  # enables log monitoring
+  logs: true
+  # enables metrics monitoring
+  metrics: true
+  # specifies output to be used
+  use_output: monitoring
+-------------------------------------------------------------------------------------
+
+
+To disable monitoring, set `settings.monitoring.enabled` to `false`. When set to
+`false`, {beats} monitoring is turned off, and all other options in this section
+are ignored.
+
+To enable monitoring, set `settings.monitoring.enabled` to `true`. Also set the
+`logs` and `metrics` settings to control whether logs, metrics, or both are
+collected. If neither setting is specified, monitoring is disabled. Set
+`use_output` to specify the output to which monitoring events are sent.
+
+[[elastic-agent-datasource-configuration]]
+== Datasource settings
+
+//TODO: provide more context here about what a data source is and how it's used.
+
+By default {agent} collects system metrics, such as cpu, memory, network, and
+filesystem metrics, and sends them to the default output. For example:
+
+
+[source,yaml]
+-------------------------------------------------------------------------------------
+datasources:
+  - namespace: default
+    use_output: default
+    inputs:
+      - type: system/metrics
+        streams:
+          - metricset: cpu
+            dataset: system.cpu
+          - metricset: memory
+            dataset: system.memory
+          - metricset: network
+            dataset: system.network
+          - metricset: filesystem
+            dataset: system.filesystem
+-------------------------------------------------------------------------------------
+
+If `use_output` is not specified, the `default` output is used.
+
+For more examples, see
+<<elastic-agent-configuration-example,`elastic-agent_configuration_example.yml`>>.