0
0
Fork 0
mirror of https://github.com/netdata/netdata.git synced 2025-04-29 23:20:01 +00:00
netdata_netdata/docs/dashboards-and-charts/anomaly-advisor-tab.md
kanelatechnical 4c96a564ee
Dashboards and charts (#20162)
Co-authored-by: Ilya Mashchenko <ilya@netdata.cloud>
2025-04-23 12:23:23 +00:00

68 lines
No EOL
3.1 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Anomaly Advisor Tab
The **Anomaly Advisor** tab helps you identify potentially anomalous metrics and charts by focusing on a highlighted timeframe of interest. This feature uses Netdata's **Anomaly Rate ML scoring** to rank metrics based on unusual behavior.
Each chart in the [Metrics tab](/docs/dashboards-and-charts/metrics-tab-and-single-node-tabs.md) also features an [Anomaly Rate ribbon](/docs/dashboards-and-charts/netdata-charts.md#anomaly-rate-ribbon) for anomaly visibility.
For configuration details, see the [ML documentation](/src/ml/README.md).
---
## How Anomaly Advisor Works
The Anomaly Advisor leverages Netdatas machine learning to evaluate anomaly rates across your nodes. It provides three key visualizations:
| Chart Name | Purpose | Why It Matters |
|--------------------------------|-------------------------------------------------------------------|----------------------------------------------------------------|
| **Anomaly Rate** | Shows the percentage of anomalous metrics over time per node. | Helps you quickly spot nodes behaving abnormally. |
| **Count of Anomalous Metrics** | Displays raw counts of anomalous metrics per node. | Useful when nodes have different numbers of collected metrics. |
| **Anomaly Events Detected** | Indicates when the anomaly rate has triggered a node-level event. | Focuses your attention on meaningful spikes, not just noise. |
:::note
**Anomaly Events Detected** appear slightly after anomaly rates rise, as they require a significant portion of metrics on the node to show anomalous behavior.
:::
---
## Workflow Overview
1. **Highlight a timeframe of interest** on the anomaly charts.
2. An ordered list of related charts appears, ranked by anomaly level.
3. The **Anomaly Rate ribbon** (purple) is visible on each chart.
4. Use the right-hand anomaly index to sort metrics from most to least anomalous.
5. Click an entry in the index to navigate directly to the corresponding chart.
:::tip
Use the [node filter](/docs/dashboards-and-charts/node-filter.md) to focus on specific nodes before highlighting a timeframe.
:::
---
## Usage Tips
| Tip | Why It Matters |
|-----------------------------------------------|------------------------------------------------------------------|
| Filter to specific nodes before highlighting. | Reduces noise by limiting averaging across unrelated nodes. |
| Highlight close to the anomaly spike. | Improves ranking accuracy by focusing on the relevant timeframe. |
---
## Anomaly Advisor Diagram
```mermaid
graph TD
A[Highlight **Timeframe**] --> B[**Rank** Metrics by Score]
B --> C[Show Ordered Charts]
C --> D[**Pick** from Anomaly Index]
D --> E[**Investigate** Metrics]
```
:::tip
This diagram shows the Anomaly Advisor flow: highlight, rank, and explore. Use the ranking to prioritize which charts to investigate.
:::