summaryrefslogtreecommitdiffstats
path: root/docs/getting-started.md
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-27 11:08:07 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-27 11:08:07 +0000
commitc69cb8cc094cc916adbc516b09e944cd3d137c01 (patch)
treef2878ec41fb6d0e3613906c6722fc02b934eeb80 /docs/getting-started.md
parentInitial commit. (diff)
downloadnetdata-upstream/1.29.3.tar.xz
netdata-upstream/1.29.3.zip
Adding upstream version 1.29.3.upstream/1.29.3upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'docs/getting-started.md')
-rw-r--r--docs/getting-started.md240
1 files changed, 240 insertions, 0 deletions
diff --git a/docs/getting-started.md b/docs/getting-started.md
new file mode 100644
index 0000000..1ccab42
--- /dev/null
+++ b/docs/getting-started.md
@@ -0,0 +1,240 @@
+<!--
+title: "Get started guide"
+date: 2020-05-04
+custom_edit_url: https://github.com/netdata/netdata/edit/master/docs/getting-started.md
+-->
+
+# Get started guide
+
+Thanks for trying the Netdata Agent! In this getting started guide, we'll quickly walk you through the first steps you
+should take after installing the Agent.
+
+The Agent can collect thousands of metrics in real-time and use its database for long-term metrics storage without any
+configuration, but there are some valuable things to know to get the most out of Netdata based on your needs.
+
+We'll skip right into some technical details, so if you're brand-new to monitoring the health and performance of systems
+and applications, our [**step-by-step guide**](/docs/guides/step-by-step/step-00.md) might be a better fit.
+
+> If you haven't installed Netdata yet, visit the [installation instructions](/packaging/installer/README.md) for
+> details, including our one-liner script, which automatically installs Netdata on almost all Linux distributions.
+
+## Access the dashboard
+
+Open up your web browser of choice and navigate to `http://NODE:19999`, replacing `NODE` with the IP address or hostname
+of your Agent. Hit **Enter**. Welcome to Netdata!
+
+![Animated GIF of navigating to the
+dashboard](https://user-images.githubusercontent.com/1153921/80825153-abaec600-8b94-11ea-8b17-1b770a2abaa9.gif)
+
+**What's next?**:
+
+- Read more about the [standard Netdata dashboard](/web/gui/).
+- Learn all the specifics of [using charts](/web/README.md#using-charts) or the differences between [charts,
+ context, and families](/web/README.md#charts-contexts-families).
+
+## Configuration basics
+
+Netdata primarily uses the `netdata.conf` file for custom configurations.
+
+On most systems, you can find that file at `/etc/netdata/netdata.conf`.
+
+> Some operating systems will place your `netdata.conf` at `/opt/netdata/etc/netdata/netdata.conf`, so check there if
+> you find nothing at `/etc/netdata/netdata.conf`.
+
+The `netdata.conf` file is broken up into various sections, such as `[global]`, `[web]`, `[registry]`, and more. By
+default, most options are commented, so you'll have to uncomment them (remove the `#`) for Netdata to recognize your
+change.
+
+Once you save your changes, [restart Netdata](#start-stop-and-restart-netdata) to load your new configuration.
+
+**What's next?**:
+
+- [Change how long Netdata stores metrics](#change-how-long-netdata-stores-metrics) by changing the `page cache size`
+ and `dbengine disk space` settings in `netdata.conf`.
+- Move Netdata's dashboard to a [different port](/web/server/) or enable TLS/HTTPS
+ encryption.
+- See all the `netdata.conf` options in our [daemon configuration documentation](/daemon/config/).
+- Run your own [registry](/registry/README.md#run-your-own-registry).
+
+## Change how long Netdata stores metrics
+
+Netdata can store long-term, historical metrics out of the box. A custom database uses RAM to store recent metrics,
+ensuring dashboards and API queries are extremely responsive, while "spilling" historical metrics to disk. This
+configuration keeps RAM usage low while allowing for long-term, on-disk metrics storage.
+
+You can tweak this custom _database engine_ to store a much larger dataset than your system's available RAM,
+particularly if you allow Netdata to use slightly more RAM and disk space than the default configuration.
+
+Read our guide on [changing how long Netdata stores metrics](/docs/store/change-metrics-storage.md) to learn more and
+use our the embedded database engine to figure out the exact settings you'll need to store historical metrics right in
+the Agent's database.
+
+**What's next?**:
+
+- Learn more about the [memory requirements for the database
+ engine](/database/engine/README.md#memory-requirements) to understand how much RAM/disk space you should commit
+ to storing historical metrics.
+
+## Collect data from more sources
+
+When Netdata _starts_, it auto-detects dozens of **data sources**, such as database servers, web servers, and more. To
+auto-detect and collect metrics from a service or application you just installed, you need to [restart
+Netdata](#start-stop-and-restart-netdata).
+
+> There is one exception: When Netdata is running on the host (as in not in a container itself), it will always
+> auto-detect containers and VMs.
+
+However, auto-detection only works if you installed the source using its standard installation procedure. If Netdata
+isn't collecting metrics after a restart, your source probably isn't configured correctly. Look at the [external plugin
+documentation](/collectors/plugins.d/) to find the appropriate module for your source. Those pages will contain
+more information about how to configure your source for auto-detection.
+
+Some modules, like `chrony`, are disabled by default and must be enabled manually for auto-detection to work.
+
+Once Netdata detects a valid source of data, it will continue trying to collect data from it. For example, if
+Netdata is collecting data from an Nginx web server, and you shut Nginx down, Netdata will collect new data as soon as
+you start the web server back up&mdash;no restart necessary.
+
+### Configure plugins
+
+Even if Netdata auto-detects your service/application, you might want to configure what, or how often, Netdata is
+collecting data.
+
+Netdata uses **internal** and **external** plugins to collect data. Internal plugins run within the Netdata dæmon, while
+external plugins are independent processes that send metrics to Netdata over pipes. There are also plugin
+**orchestrators**, which are external plugins with one or more data collection **modules**.
+
+You can configure both internal and external plugins, along with the individual modules. There are many ways to do so:
+
+- In `netdata.conf`, `[plugins]` section: Enable or disable internal or external plugins with `yes` or `no`.
+- In `netdata.conf`, `[plugin:XXX]` sections: Each plugin has a section for changing collection frequency or passing
+ options to the plugin.
+- In `.conf` files for each external plugin: For example, at `/etc/netdata/python.d.conf`.
+- In `.conf` files for each module : For example, at `/etc/netdata/python.d/nginx.conf`.
+
+It's complex, so let's walk through an example of the various `.conf` files responsible for collecting data from an
+Nginx web server using the `nginx` module and the `python.d` plugin orchestrator.
+
+First, you can enable or disable the `python.d` plugin entirely in `netdata.conf`.
+
+```conf
+[plugins]
+ # Enabled
+ python.d = yes
+ # Disabled
+ python.d = no
+```
+
+You can also configure the entire `python.d` external plugin via the `[plugin:python.d]` section in `netdata.conf`.
+Here, you can change how often Netdata uses `python.d` to collect metrics or pass other command options:
+
+```conf
+[plugin:python.d]
+ update every = 1
+ command options =
+```
+
+The `python.d` plugin has a separate configuration file at `/etc/netdata/python.d.conf` for enabling and disabling
+modules. You can use the `edit-config` script to edit the file, or open it with your text editor of choice:
+
+```bash
+sudo /etc/netdata/edit-config python.d.conf
+```
+
+Finally, the `nginx` module has a configuration file called `nginx.conf` in the `python.d` folder. Again, use
+`edit-config` or your editor of choice:
+
+```bash
+sudo /etc/netdata/edit-config python.d/nginx.conf
+```
+
+In the `nginx.conf` file, you'll find additional options. The default works in most situations, but you may need to make
+changes based on your particular Nginx setup.
+
+**What's next?**:
+
+- Look at the [full list of data collection modules](/collectors/COLLECTORS.md)
+ to configure your sources for auto-detection and monitoring.
+- Improve the [performance](/docs/guides/configure/performance.md) of Netdata on low-memory systems.
+- Configure `systemd` to expose [systemd services
+ utilization](/collectors/cgroups.plugin/README.md#monitoring-systemd-services) metrics automatically.
+- [Reconfigure individual charts](/daemon/config/README.md#per-chart-configuration) in `netdata.conf`.
+
+## Health monitoring and alarms
+
+Netdata comes with hundreds of health monitoring alarms for detecting anomalies on production servers. If you're running
+Netdata on a workstation, you might want to disable Netdata's alarms.
+
+Edit your `/etc/netdata/netdata.conf` file and set the following:
+
+```conf
+[health]
+ enabled = no
+```
+
+If you want to keep health monitoring enabled, but turn email notifications off, edit your `health_alarm_notify.conf`
+file with `edit-config`, or with the text editor of your choice:
+
+```bash
+sudo /etc/netdata/edit-config health_alarm_notify.conf
+```
+
+Find the `SEND_EMAIL="YES"` line and change it to `SEND_EMAIL="NO"`.
+
+**What's next?**:
+
+- Follow the [health quickstart](/health/QUICKSTART.md) to locate and edit existing health entities, and then
+ create your own.
+- See all the alarm options via the [health configuration reference](/health/REFERENCE.md).
+- Add a new notification method, like [Slack](/health/notifications/slack/).
+
+## Monitor multiple systems with Netdata Cloud
+
+If you have the Agent installed on multiple nodes, you can use Netdata Cloud in two ways: Monitor the health and
+performance of an entire infrastructure via the Netdata Cloud web application, or use the Visited Nodes menu that's
+built into every dashboard.
+
+![The War Room
+Overview](https://user-images.githubusercontent.com/1153921/102651377-b1f4b100-4129-11eb-8e60-d2995d258c16.png)
+
+You can use these features together or separately&mdash;the decision is up to you and the needs of your infrastructure.
+
+**What's next?**:
+
+- Sign up for [Netdata Cloud](https://app.netdata.cloud).
+- Read the [infrastructure monitoring quickstart](/docs/quickstart/infrastructure.md).
+- Better understand how the Netdata Agent connects securely to Netdata Cloud with [claiming](/claim/README.md) and
+ [Agent-Cloud link](/aclk/README.md) documentation.
+
+## Start, stop, and restart Netdata
+
+When you install Netdata, it's configured to start at boot, and stop and restart/shutdown. You shouldn't need to start
+or stop Netdata manually, but you will probably need to restart Netdata at some point.
+
+- To **start** Netdata, open a terminal and run `service netdata start`.
+- To **stop** Netdata, run `service netdata stop`.
+- To **restart** Netdata, run `service netdata restart`.
+
+The `service` command is a wrapper script that tries to use your system's preferred method of starting or stopping
+Netdata based on your system. But, if either of those commands fails, try using the equivalent commands for `systemd`
+and `init.d`:
+
+- **systemd**: `systemctl start netdata`, `systemctl stop netdata`, `systemctl restart netdata`
+- **init.d**: `/etc/init.d/netdata start`, `/etc/init.d/netdata stop`, `/etc/init.d/netdata restart`
+
+## What's next?
+
+Even after you've configured `netdata.conf`, tweaked alarms, learned the basics of performance troubleshooting, and
+claimed all your systems in Netdata Cloud or added them to the Visited nodes menu, you've just gotten started with
+Netdata.
+
+Take a look at some more advanced features and configurations:
+
+- Centralize Netdata metrics from many systems with [streaming](/streaming/README.md)
+- Enable long-term archiving of Netdata metrics via [exporting engine](/exporting/README.md) to time-series databases.
+- Improve security by putting Netdata behind an [Nginx proxy with SSL](/docs/Running-behind-nginx.md).
+
+Or, learn more about how you can contribute to [Netdata core](/CONTRIBUTING.md) or our
+[documentation](/docs/contributing/contributing-documentation.md)!
+
+[![analytics](https://www.google-analytics.com/collect?v=1&aip=1&t=pageview&_s=1&ds=github&dr=https%3A%2F%2Fgithub.com%2Fnetdata%2Fnetdata&dl=https%3A%2F%2Fmy-netdata.io%2Fgithub%2Fdocs%2Fgetting-started&_u=MAC~&cid=5792dfd7-8dc4-476b-af31-da2fdb9f93d2&tid=UA-64295674-3)](<>)