master
md 303 lines 11.5 KB
Rendered Raw
1 # Reporting
2
3 Extract infrastructure insights for stakeholders, leadership, and external tools. Netdata provides four ways to create reports—from asking a simple question to exporting raw metrics into your existing business intelligence stack.
4
5 ## Methods overview
6
7 | Method | Effort | Best for |
8 |--------|--------|----------|
9 | **[AI Insights](#ai-insights)** | Easiest | Executive summaries, recurring reports, natural language queries |
10 | **[AI Assistants (MCP)](#ai-assistants-mcp)** | Easy | Ad-hoc analysis, deep investigation, developer workflows |
11 | **[Grafana](#grafana-integration)** | Medium | Custom dashboards, teams already using Grafana |
12 | **[Export to BI](#export-to-business-intelligence-tools)** | Advanced | Power BI, Tableau, Looker, custom analytics pipelines |
13
14 ## AI Insights
15
16 Ask Netdata anything about your infrastructure in plain language and receive an executive-ready report. No configuration required—just describe what you need.
17
18 ### When to use it
19
20 - Monday morning recap of weekend incidents
21 - Post-incident executive summary for leadership
22 - Weekly health checks and situational awareness
23 - Cost optimization analysis
24 - SLO conformance reporting
25
26 ### How to generate a report
27
28 1. Open Netdata Cloud and go to `Insights`
29 2. Select a pre-built report type (Infrastructure Summary, Performance Optimization, Capacity Planning, Anomaly Analysis) or click `New Investigation` for a custom prompt
30 3. Choose the time range and scope (all nodes or a specific room/space)
31 4. Click `Generate`
32
33 Reports complete in 2–3 minutes and are saved in Insights. You receive an email when the report is ready.
34
35 ### Scheduling recurring reports
36
37 Automate your reporting workflow with scheduled reports:
38
39 1. Configure your report as above
40 2. Click `Schedule` (next to `Generate`)
41 3. Choose cadence: daily, weekly, or monthly
42 4. Set the delivery time
43
44 Scheduled reports run automatically and deliver results to your email and the Insights tab.
45
46 ### Example prompts
47
48 **Weekly infrastructure health:**
49 ```
50 Generate a weekly infrastructure summary for services A, B, and C.
51 Include major incidents, anomalies, capacity risks, and recommended follow-ups.
52 ```
53
54 **Cost optimization:**
55 ```
56 Identify underutilized nodes for cost savings. Monthly compute is ~$12K
57 with mixed workloads. Goal: save $2–3K/month without reliability impact.
58 ```
59
60 **SLO conformance:**
61 ```
62 Generate an SLO conformance report for 'user-auth' (99.9% uptime,
63 p95 latency <200ms) for the last 7 days. Include breaches, contributing
64 factors, and remediation recommendations.
65 ```
66
67 ### Availability
68
69 - Available to Business and Free Trial plans
70 - Each report consumes 1 AI credit (10 free per month on eligible plans)
71 - Data privacy: metrics are summarized into structured context; your data is not used to train foundation models
72
73 ## AI Assistants (MCP)
74
75 Connect your AI assistant directly to Netdata using the Model Context Protocol (MCP). Ask questions in natural language and receive answers based on live infrastructure data.
76
77 MCP is available in two ways:
78
79 - **Netdata Cloud MCP** at `app.netdata.cloud/api/v1/mcp` — infrastructure-wide access to all your nodes (Paid plan)
80 - **Agent/Parent MCP** on every Netdata Agent and Parent (v2.6.0+) — direct local access (free, open-source)
81
82 AI assistants can query metrics, alerts, logs, and live system information across your entire infrastructure.
83
84 ### Supported AI clients
85
86 | Client | Description |
87 |--------|-------------|
88 | Claude Desktop | Anthropic's desktop AI assistant |
89 | Claude Code | Anthropic's CLI for development workflows |
90 | Cursor | AI-powered code editor |
91 | VS Code | Visual Studio Code with MCP support |
92 | JetBrains IDEs | IntelliJ, PyCharm, WebStorm, and others |
93 | Gemini CLI | Google's Gemini CLI |
94 | OpenAI Codex CLI | OpenAI's development tools |
95
96 ### How to connect
97
98 ```bash
99 # Export your MCP key
100 export NETDATA_MCP_API_KEY="$(cat /var/lib/netdata/mcp_dev_preview_api_key)"
101
102 # Connect using mcp-remote
103 npx mcp-remote@latest --http http://YOUR_NETDATA_IP:19999/mcp \
104 --allow-http \
105 --header "Authorization: Bearer $NETDATA_MCP_API_KEY"
106 ```
107
108 ### Example queries
109
110 Once connected, ask natural language questions:
111
112 - "Show me CPU usage for all nodes in the last 24 hours"
113 - "What are the top 10 processes by memory consumption?"
114 - "Find any anomalies in network traffic this week"
115 - "Generate a summary of all active alerts"
116 - "Which nodes have the highest disk utilization?"
117
118 ### Availability
119
120 - **Cloud MCP**: Available on Paid plans — infrastructure-wide access, zero local setup
121 - **Agent/Parent MCP**: Available on all plans (v2.6.0+) — free, open-source, direct local access
122 - Unlimited queries with no per-query charges
123 - Requires API key (Cloud token or local API key) for access
124
125 See [Netdata MCP](/docs/netdata-ai/mcp/README.md) for detailed setup instructions.
126
127 ## Grafana integration
128
129 Connect Grafana to Netdata Cloud for infrastructure-wide dashboards. Use Grafana's visualization capabilities with Netdata's real-time metrics.
130
131 ### When to use it
132
133 - Teams already using Grafana for other data sources
134 - Custom dashboard requirements beyond Netdata's built-in charts
135 - Combining Netdata metrics with data from other systems
136
137 ### How to connect
138
139 1. Install the Netdata data source plugin in Grafana
140 2. Configure connection to Netdata Cloud using an API token
141 3. Create dashboards using Grafana's query builder
142
143 :::tip
144
145 Generate API tokens from Netdata Cloud under **User Settings****API Tokens**. See [API Tokens](/docs/netdata-cloud/authentication-and-authorization/api-tokens.md) for details.
146
147 :::
148
149 ### Availability
150
151 - Requires Netdata Cloud account
152 - Grafana Cloud or self-hosted Grafana
153 - API tokens available on all plans
154
155 ## Export to business intelligence tools
156
157 Export metrics from Netdata to external databases and business intelligence platforms. You can query data from individual Agents or use Netdata Cloud to aggregate metrics from your entire infrastructure.
158
159 ### Supported BI platforms
160
161 Netdata integrates with popular business intelligence tools through several pathways:
162
163 | BI Platform | Integration Options |
164 |-------------|---------------------|
165 | **Power BI** | Netdata Cloud API, Prometheus endpoint, or database export |
166 | **Tableau** | Netdata Cloud API, PostgreSQL, or Prometheus |
167 | **Looker / Looker Studio** | Netdata Cloud API, BigQuery, or Prometheus |
168 | **Qlik** | Netdata Cloud API, PostgreSQL, or InfluxDB |
169 | **SAP Analytics Cloud** | Netdata Cloud API or PostgreSQL |
170 | **Metabase** | Netdata Cloud API, PostgreSQL, or TimescaleDB |
171 | **Apache Superset** | Netdata Cloud API, PostgreSQL, or Prometheus |
172 | **Domo** | Netdata Cloud API or database connectors |
173 | **ThoughtSpot** | Netdata Cloud API or PostgreSQL |
174
175 ### Query options
176
177 #### Netdata Cloud API (recommended)
178
179 The Netdata Cloud API lets you query metrics from all your nodes through a single endpoint. This is the simplest approach for multi-node infrastructure.
180
181 1. Generate an API token from **User Settings****API Tokens**
182 2. Use the token to authenticate requests to `https://app.netdata.cloud/api/v2/data`
183
184 ```bash
185 # Query CPU metrics from all nodes
186 curl -H 'Accept: application/json' \
187 -H "Authorization: Bearer YOUR_API_TOKEN" \
188 'https://app.netdata.cloud/api/v2/data?contexts=system.cpu&after=-3600'
189
190 # Get list of all nodes in your space
191 curl -H 'Accept: application/json' \
192 -H "Authorization: Bearer YOUR_API_TOKEN" \
193 'https://app.netdata.cloud/api/v2/nodes'
194 ```
195
196 The Cloud API returns aggregated data from all nodes in your infrastructure, making it ideal for BI tools that need a unified view.
197
198 #### Prometheus endpoint (single-node)
199
200 For single-node deployments or Prometheus-based workflows, query the Agent or Parent directly:
201
202 ```
203 http://NODE_IP:19999/api/v3/allmetrics?format=prometheus
204 ```
205
206 Replace `NODE_IP` with your Netdata Agent or Parent IP address:
207 - **Agent IP**: Returns metrics from that single node
208 - **Parent IP**: Returns aggregated metrics from all child agents connected to that Parent
209
210 This endpoint is useful when you need metrics from a specific node or when your BI tool already integrates with Prometheus.
211
212 #### REST API with JSON
213
214 Query specific metrics from an Agent or Parent in JSON format. This is useful for BI tools that need to combine Netdata metrics with other business data.
215
216 **Common BI use cases:**
217
218 ```bash
219 # Daily averages for last 30 days, grouped by node
220 curl 'http://NODE_IP:19999/api/v3/data?contexts=system.cpu&after=-2592000&points=30&time_group=avg&group_by=node'
221
222 # Weekly max values for capacity planning
223 curl 'http://NODE_IP:19999/api/v3/data?contexts=system.ram&after=-604800&points=4&time_group=max&group_by=node'
224
225 # Hourly sum for cost analysis
226 curl 'http://NODE_IP:19999/api/v3/data?contexts=system.cpu&after=-86400&points=24&time_group=sum'
227 ```
228
229 **Key parameters for BI workflows:**
230
231 | Parameter | Description | Example |
232 |-----------|-------------|---------|
233 | `contexts` | Metric context to query | `system.cpu`, `system.ram`, `disk.io` |
234 | `after` / `before` | Timeframe (seconds or Unix timestamp) | `-2592000` = last 30 days |
235 | `points` | Number of output points | `30` = daily for monthly view |
236 | `time_group` | Aggregation function | `avg`, `sum`, `min`, `max` |
237 | `group_by` | How to group results | `node`, `context`, `label:LABEL_NAME` |
238
239 Power BI, Tableau, and similar tools can consume this JSON through their data transformation features (Power Query, etc.).
240
241 #### Database export connectors
242
243 For persistent storage and historical analysis, export metrics to a database:
244
245 | Database | Connector |
246 |----------|-----------|
247 | PostgreSQL | Prometheus remote write adapter |
248 | TimescaleDB | Prometheus remote write or netdata-timescale-relay |
249 | InfluxDB | Graphite or Prometheus remote write |
250 | Elasticsearch | Graphite or Prometheus remote write |
251 | Google BigQuery | Prometheus remote write |
252 | AWS services | AWS Kinesis Data Streams |
253 | Azure services | Prometheus remote write |
254
255 See [Export Metrics to External Time-Series Databases](/docs/exporting-metrics/README.md) for full connector documentation.
256
257 ### When to use this approach
258
259 - Existing BI infrastructure with established workflows
260 - Requirement to combine Netdata data with other business data sources
261 - Custom visualization needs not covered by built-in dashboards
262 - Long-term archival beyond Netdata's retention
263 - Compliance requirements for data export
264
265 ## Choosing the right method
266
267 ### Start with AI Insights if you:
268
269 - Need reports quickly without setup
270 - Want executive-ready summaries
271 - Prefer natural language over configuration
272 - Need recurring automated reports
273
274 ### Use AI Assistants (MCP) if you:
275
276 - Want real-time answers to ad-hoc questions
277 - Already use Claude, Cursor, or similar AI tools
278 - Need deep investigation capabilities
279 - Prefer conversational interaction with your data
280
281 ### Use Grafana if you:
282
283 - Already have Grafana deployed
284 - Need highly customized dashboards
285 - Want to combine Netdata with other data sources in one view
286 - Have team expertise in Grafana
287
288 ### Export to BI tools if you:
289
290 - Have established Power BI, Tableau, or Looker workflows
291 - Need to combine infrastructure metrics with business data
292 - Require custom analytics beyond monitoring
293 - Have compliance requirements for data in specific systems
294
295 ## Related documentation
296
297 - [AI Insights: Infrastructure Summary](/docs/netdata-ai/insights/infrastructure-summary.md)
298 - [AI Insights: Scheduled Reports](/docs/netdata-ai/insights/scheduled-reports.md)
299 - [Investigations](/docs/netdata-ai/investigations/index.md)
300 - [Netdata MCP](/docs/netdata-ai/mcp/README.md)
301 - [API Tokens](/docs/netdata-cloud/authentication-and-authorization/api-tokens.md)
302 - [Export Metrics to External Time-Series Databases](/docs/exporting-metrics/README.md)
303 - [REST API](/src/web/api/README.md)