master
md 314 lines 13.5 KB
Rendered Raw
1 # Expanded Chart Analysis
2
3 Analyze your metrics in-depth with Netdata's Expanded Chart Analysis feature. Access detailed statistics, discover correlated metrics, and investigate anomalies directly from any chart.
4
5 ## Overview
6
7 The Expanded Chart Analysis feature provides an analytical panel below your charts. It gives you four ways to investigate your data: view detailed statistics, explore hierarchical relationships, compare time periods, and discover correlated metrics across your infrastructure.
8
9 No query language required—everything works with point-and-click.
10
11 ## Open the Analysis Panel
12
13 Click the **Expand** button at the bottom of any chart. The analysis panel opens below the chart with four analysis modes available.
14
15 ![expand button](https://raw.githubusercontent.com/netdata/docs-images/7a86c33450c24df340b7381c2c9587f332ad0c04/expanded-analysis/Expand%20Chart%20Analysis.png)
16
17 When you're done, click **Collapse** to close the panel. Your settings persist and restore when you reopen it.
18
19 ## Select Your Time Range
20
21 Choose which time period to analyze using the tabs at the top of the panel:
22
23 **Window** analyzes the entire visible time range on the chart. This tab is always available and updates automatically when you zoom or pan the chart.
24
25 **Selected Area** analyzes a specific highlighted period. This tab activates when you highlight a time range on the chart (see [Highlight a Time Range](#highlight-a-time-range)).
26
27 All analysis modes support both time range options.
28
29 ## Analysis Modes
30
31 Switch between modes by clicking the icons in the panel header. Each mode is designed for a specific type of investigation.
32
33 | Mode | Icon | Use For |
34 |-----------------------|------|----------------------------------------------------|
35 | **Chart Values** | 📊 | Viewing detailed statistics for all dimensions |
36 | **Drill Down** | 🔽 | Exploring which nodes or instances contribute most |
37 | **Compare Periods** | 📈 | Comparing current metrics with historical data |
38 | **Correlate Metrics** | 🔗 | Finding related metrics across your infrastructure |
39
40 ### Chart Values
41
42 View detailed statistics for every dimension in your chart. This mode shows you the exact numbers behind the visualization.
43
44 ![chart values](https://raw.githubusercontent.com/netdata/docs-images/7a86c33450c24df340b7381c2c9587f332ad0c04/expanded-analysis/Chart%20Values.png)
45
46 | Column | Description |
47 |---------------|------------------------------------------------------|
48 | **Dimension** | Dimension name with color indicator |
49 | **Value** | Current or hovering value with units |
50 | **Min** | Minimum value in the time period |
51 | **Avg** | Average value in the time period |
52 | **Max** | Maximum value in the time period |
53 | **Anomaly** | Anomaly rate (percentage of time behaving unusually) |
54
55 :::tip
56
57 Sort by the Anomaly column to quickly find which dimensions are behaving strangely.
58
59 :::
60
61 **Show advanced statistics**
62
63 Click the metrics icon in the panel header to reveal additional columns:
64
65 | Advanced Column | Description |
66 |-----------------|--------------------------------|
67 | **Median** | Middle value (50th percentile) |
68 | **StdDev** | Standard deviation |
69 | **P95** | 95th percentile |
70 | **Range** | Difference between max and min |
71 | **Volume** | Sum of all values |
72
73 **Sort and export**
74
75 Click any column header to sort by that column. Click **Download** to export the data as CSV.
76
77 **When to use this mode:**
78
79 Use Chart Values when investigating spikes to see which specific dimensions are responsible. Sort by Anomaly to find unusual behavior, or check Max values to see peak usage.
80
81 ### Drill Down
82
83 Explore hierarchical relationships in your infrastructure. This mode organizes metrics into a three-level tree:
84
85 ![drill down](https://raw.githubusercontent.com/netdata/docs-images/7a86c33450c24df340b7381c2c9587f332ad0c04/expanded-analysis/Drill%20Down.png)
86
87 ```mermaid
88 flowchart TD
89 A("**Node Level**<br/>server-prod-01<br/>Contribution: 65%") --> B("**Instance Level**<br/>disk: sda<br/>Contribution: 45%")
90 A --> C("**Instance Level**<br/>disk: sdb<br/>Contribution: 20%")
91 B --> D("**Dimension Level**<br/>read<br/>Contribution: 30%")
92 B --> E("**Dimension Level**<br/>write<br/>Contribution: 15%")
93 C --> F("**Dimension Level**<br/>read<br/>Contribution: 12%")
94 C --> G("**Dimension Level**<br/>write<br/>Contribution: 8%")
95 H("**Node Level**<br/>server-prod-02<br/>Contribution: 35%") --> I("**Instance Level**<br/>disk: sda<br/>Contribution: 35%")
96 I --> J("**Dimension Level**<br/>read<br/>Contribution: 20%")
97 I --> K("**Dimension Level**<br/>write<br/>Contribution: 15%")
98 %% Style definitions
99 classDef alert fill: #ffeb3b, stroke: #000000, stroke-width: 3px, color: #000000, font-size: 18px
100 classDef neutral fill: #f9f9f9, stroke: #000000, stroke-width: 3px, color: #000000, font-size: 18px
101 classDef complete fill: #4caf50, stroke: #000000, stroke-width: 3px, color: #000000, font-size: 18px
102 classDef database fill: #2196F3, stroke: #000000, stroke-width: 3px, color: #000000, font-size: 18px
103 %% Apply styles
104 class A alert
105 class H alert
106 class B database
107 class C database
108 class I database
109 class D neutral
110 class E neutral
111 class F neutral
112 class G neutral
113 class J neutral
114 class K neutral
115 ```
116
117 | Level | Description | Examples |
118 |---------------|-----------------------------------|----------------------------|
119 | **Node** | Individual servers or hosts | server-01, web-prod-03 |
120 | **Instance** | Specific entities being monitored | disk "sda", network "eth0" |
121 | **Dimension** | Individual metrics | "read", "write", "used" |
122
123 Each level shows:
124
125 | Column | Description |
126 |------------------|----------------------------------------------------|
127 | **Name** | Node, instance, or dimension name with expand icon |
128 | **Contribution** | Percentage weight (contribution to total) |
129 | **Anomaly Rate** | Aggregated anomaly rate for this branch |
130 | **Min** | Minimum value in period |
131 | **Avg** | Average value in period |
132 | **Max** | Maximum value in period |
133
134 **Navigate the hierarchy**
135
136 Click the expand icon next to any row to drill into the next level. Click again to collapse. Click **Group By** to change how the hierarchy is organized.
137
138 :::info
139
140 The contribution percentage shows the weight of each item relative to the total. A node contributing 80% tells you that most of your metrics come from that server.
141
142 :::
143
144 **When to use this mode:**
145
146 Use Drill Down to identify which node is using the most resources. Expand the Node level and sort by Contribution to see which servers contribute most. Click the node name to navigate to that node's dashboard.
147
148 ### Compare Periods
149
150 Compare current metrics with historical periods to identify trends and changes. This mode automatically shows comparisons with:
151
152 ![compare periods](https://raw.githubusercontent.com/netdata/docs-images/7a86c33450c24df340b7381c2c9587f332ad0c04/expanded-analysis/Compare.png)
153
154 | Period | Description |
155 |----------------|--------------------------------------------------------|
156 | **Current** | The visible time window or selected area (base period) |
157 | **Yesterday** | Same duration, 24 hours ago |
158 | **Last Week** | Same duration, 7 days ago |
159 | **Last Month** | Same duration, 30 days ago |
160
161 Each period card displays:
162
163 - Period label and date range
164 - Basic statistics (Min, Avg, Max)
165 - Change indicators showing percentage and absolute change compared to current
166
167 | Indicator | Meaning |
168 |-----------|------------------------------|
169 | 🔴 Red | Increase from current period |
170 | 🟢 Green | Decrease from current period |
171 | ⚪ Gray | No significant change |
172
173 **Show advanced statistics**
174
175 Click the metrics icon in the panel header to reveal Median, StdDev, P95, Range, and Volume for each period. All statistics include change indicators.
176
177 **Add custom periods**
178
179 Compare with specific dates important to your analysis. Click **Add Custom Period**, enter a label (e.g., "Black Friday 2024"), set the start and end date/time, and click **Add**. The custom period appears in the comparison grid.
180
181 Click the edit icon on any custom period card to modify the label or date range.
182
183 :::warning
184
185 Historical data might not be available if it exceeds your retention settings. Use shorter comparison periods if you see errors.
186
187 :::
188
189 **When to use this mode:**
190
191 Use Compare Periods to identify growth trends over time. Review Avg values across all periods to spot consistent increases or decreases. Export the data for capacity planning.
192
193 ### Correlate Metrics
194
195 Discover which metrics across your infrastructure correlate with the current chart's behavior. This mode shows you metrics that follow similar patterns, helping you find related issues.
196
197 ![correlate metrics](https://raw.githubusercontent.com/netdata/docs-images/7a86c33450c24df340b7381c2c9587f332ad0c04/expanded-analysis/Correlate.png)
198
199 The table displays correlated metrics grouped by context:
200
201 | Column | Description |
202 |---------------|---------------------------------------------------|
203 | **Context** | Metric category (grouped header) |
204 | **Metric** | Full metric name (node:instance:dimension) |
205 | **Weight** | Correlation weight (lower = stronger correlation) |
206 | **Sparkline** | Mini chart showing the metric's behavior |
207
208 **Choose your correlation method**
209
210 Select how to calculate correlations:
211
212 | Method | Description | Best For |
213 |------------|-----------------------------------------------------|------------------------------|
214 | **Volume** | Identifies metrics with similar magnitude changes | Most use cases (recommended) |
215 | **KS2** | Uses statistical analysis for similar distributions | Complex pattern matching |
216
217 **Select aggregation**
218
219 Choose how to aggregate data:
220
221 | Aggregation | Use When |
222 |-------------|--------------------------------|
223 | **Average** | General patterns (recommended) |
224 | **Median** | Outliers might skew results |
225 | **Min** | Focusing on minimum values |
226 | **Max** | Focusing on peak values |
227 | **StdDev** | Focusing on variability |
228
229 **Adjust the threshold**
230
231 Use the threshold slider to control how many results appear. Lower threshold (e.g., 0.1%) shows more results including weaker correlations. Higher threshold (e.g., 5%) shows fewer results with only the strongest correlations.
232
233 :::tip
234
235 Start with 1-2% threshold for most investigations. Increase the threshold if you see too many results, or decrease it if you need more options.
236
237 :::
238
239 **Navigate to correlated metrics**
240
241 Click any metric row to navigate to that chart and investigate further.
242
243 **When to use this mode:**
244
245 Use Correlate Metrics when investigating a spike or anomaly. Highlight the problem period, switch to Selected Area tab, and review correlated metrics to find related issues across your infrastructure.
246
247 ## Highlight a Time Range
248
249 Focus your analysis on a specific time period by highlighting it on the chart.
250
251 Hold **Alt** (Windows/Linux) or **⌘** (macOS), then click and drag across the chart to select a time range. The highlighted area appears with a visual overlay.
252
253 Once you create a highlight, the **Selected Area** tab becomes available in the analysis panel. Click it to view statistics and correlations for only the highlighted period.
254
255 To remove a highlight, click anywhere on the chart outside the highlighted area.
256
257 :::note
258
259 The **Selected Area** tab is disabled until you create a highlight on the chart.
260
261 :::
262
263 ## Troubleshooting
264
265 <details>
266 <summary><strong>The Selected Area tab is grayed out</strong></summary>
267
268 You need to highlight a time range on the chart first. Hold **Alt** (or **⌘** on macOS) and drag across the chart to create a highlight.
269
270 </details>
271
272 <details>
273 <summary><strong>No correlated metrics appear</strong></summary>
274
275 The threshold might be set too low. Try increasing the threshold slider to 1-5% to see the strongest correlations.
276
277 </details>
278
279 <details>
280 <summary><strong>Drill Down shows "No data available"</strong></summary>
281
282 This occurs when:
283
284 - The time range is too short
285 - No data was collected during this period
286
287 Try zooming out to a longer time range on the chart.
288
289 </details>
290
291 <details>
292 <summary><strong>Compare mode shows errors for historical periods</strong></summary>
293
294 Historical data might not be available due to retention settings. Try using shorter comparison periods (Yesterday instead of Last Month) or check your data retention configuration.
295
296 </details>
297
298 <details>
299 <summary><strong>Advanced statistics columns are missing</strong></summary>
300
301 Click the metrics icon button in the panel header to show advanced statistics (Median, StdDev, P95, Range, Volume).
302
303 </details>
304
305 <details>
306 <summary><strong>The analysis panel is slow to load</strong></summary>
307
308 For large time ranges or many dimensions:
309
310 - Use a shorter time range
311 - Highlight a specific area to analyze
312 - Filter to fewer dimensions using the chart legend
313
314 </details>