One plot, every chart typeSame diaries, different pictures
The data never changes here: four MTS of cpu.utilization, one per host. Switch the visualization and notice what each type keeps and what it throws away.
This lesson covers how you find a metric, plot it in Chart Builder, pick a chart type, and read the result correctly. It also covers how charts live in dashboards and dashboard groups, how built-in dashboards differ from custom ones, and how to subscribe to alerts. Reading charts correctly means knowing how chart resolution, rollups and analytics functions change what you see.
Think of Splunk Observability Cloud as a security-camera room. Metrics are the cameras, a chart is one monitor, and a dashboard is the wall of monitors. This lesson teaches you how to find the right camera, choose the right kind of monitor, read the picture without being fooled by blur, and sign up for a text when something goes wrong.
The data never changes here: four MTS of cpu.utilization, one per host. Switch the visualization and notice what each type keeps and what it throws away.
The Metric Finder searches metric names (full or partial), metadata (dimensions, properties, tags), metadata values, integration names and environment properties. Typing a dot shows completions for a prefix. Typing a colon after a dimension or property name lists its values.
The Metadata Catalog (Settings, then Metric Metadata) searches metrics, dimensions, custom properties and tags. In Chart Builder itself, the Signal field takes wildcards and the Browse sidebar has Find metrics.
The Metric Finder is the search bar for the camera room. Type part of a name, a host, or a tag and it shows every camera that matches. It is a simple search: no AND or OR, no fancy wildcards in plain text.
The Metadata Catalog is the phone book: look up a dimension like host.name and see which metrics carry it.
Open Chart Builder from Dashboards, then Create (+) > Chart. Each metric you enter in the Signal field becomes a plot (plot line) with a letter. You can enter a metric name that is not sending data yet.
(A/(A+B)) * 100. Hide the source plots with the eye icon to show only the result.Save with Save, Save as or Save and close. Closing a chart does not prompt you, and unsaved changes are lost.
Chart Builder is the workshop where you build one monitor. Each metric you add is a colored line called a plot. You can do math between lines with a formula, like A divided by (A plus B) times 100 for a percentage.
Careful: closing the workshop does not ask to save. Save first or your work is gone.
| Where | What it controls |
|---|---|
| Plot Editor / plot line | One metric or formula, its analytics, name, visibility, axis |
| Plot configuration (gear) | Units, per-plot visualization type, color, rollup, extrapolation |
| Chart Options tab | Whole-chart settings: minimum resolution, max delay, stack, sort, color by value |
| Data Table tab | Raw point values, columns, CSV export |
| Events tab | Custom events and alert markers |
Graph charts show points over time and come as line, area, column and histogram. Other types: pie/donut, list, single value, heatmap, event feed, text, table and service map.
Color by value applies to single value and list charts, stack applies to area and column, sort applies to list charts.
Pick the monitor that fits the question.
| Type | Best for | Remember |
|---|---|---|
| Line / Area / Column | Trend over time | Area fills below the line, column can stack |
| Histogram | Distribution of values | Can group in two levels |
| List | Ranked current values | Max 100 items |
| Single value | One headline number | First MTS only |
| Heatmap | Spot outlier sources | Group by up to 2 dimensions |
| Event feed | Events | Not metric data |
| Text | Notes, instructions | Markdown, no images |
| Table | Metrics and dimensions in rows/columns | One row per MTS |
Chart resolution is the interval at which points appear on a chart. It is separate from the native resolution at which data was sent. Short time ranges usually show native resolution, long ranges are coarser. A chart has one resolution shared by all plots, and when plots differ the chart uses the coarsest one.
A rollup reduces all points for an MTS in an interval to one point. Rollups run after data is retrieved from storage and before analytics functions. Defaults: Average for gauges, Sum for counters. Others: Min, Max, Count, Latest, Lag, Rate/sec, Delta (for cumulative counters, never negative).
Analytics: an aggregation combines MTS at each point in time (one line). A transformation works over a moving or calendar window of one MTS (for example a 5m moving mean). Timeshift plus a formula shows change versus an earlier time.
Resolution is how zoomed-in the picture is. If the camera records every 10 seconds but you view a month, the screen cannot show every frame, so it squishes many frames into one dot. The squish rule is the rollup.
Aggregation means blend many lines into one. Transformation means smooth one line over time, like a moving average.
| Metric type / goal | Rollup |
|---|---|
| Gauge (default) | Average |
| Counter (default) | Sum |
| Cumulative counter change | Delta |
| Keep spikes | Max |
| Keep dips | Min |
| Counter per second | Rate/sec |
| Value of last point | Latest |
A dashboard is a grouping of charts, and every dashboard belongs to a dashboard group. Create a dashboard from Create (+) > Dashboard, and a group from Create > Dashboard Group (this makes a custom group). Setting permissions on groups is an Enterprise Edition feature.
! excludes.A dashboard is a wall of monitors. A dashboard group is the room that holds several walls. Every wall must be in a room.
| Type | Created | Editable? |
|---|---|---|
| Built-in | Automatically when you add a data-monitoring integration | Read-only, use Save as |
| Custom | Manually, by creating a dashboard group | Yes, shared work |
| User | Automatically per user after first login | Personal workspace |
Built-in dashboards give immediate visibility into the technologies in your environment. They are created automatically when you deploy a supported integration, so you need the integration to get them. The dashboards and charts are read-only: to change one, clone the dashboard (Dashboard actions > Save dashboard as...) or clone a chart (open it, then Save as...). Find them under Dashboards, in the Built-in dashboard groups section, or in a navigator.
Subscribe in two ways: Alerts > Detectors, the more icon, Manage Subscriptions, Add Recipient; or the bell icon on a chart with linked detectors, choose the detector, then Subscribe. Recipients are set per rule. Email To, Cc and Bcc share a limit of 50 recipients. If notifications are missing, check for deactivated rules and the Muting Rules tab.
Built-in dashboards are pre-made walls of monitors that appear when you connect a data source. They are locked. To customize, make a copy with Save as and edit the copy.
To get on the call list: either open the detector's menu and pick Manage Subscriptions, or click the bell on a chart, pick the detector, and press Subscribe. Each alarm rule has its own list. If nobody gets texts, check whether a rule is turned off or a muting rule is active.
Uses lab.cpu.utilization from send.sh mts. Built-in dashboards appear only once an integration sends data, so one step uses your org's Organization Overview instead.
Questions are worded like the real exam. Press 1 to 4 to answer and Enter to continue.
Questions are worded like the real exam. Read the explanation after each one, even when you get it right.