The Scenario
The infrastructure architect on your team has been asked to put together a hardware-scaling proposal for the next fiscal year. The ask came from the VP of Engineering on a Tuesday afternoon: a one-time snapshot of the 12-node Elasticsearch cluster — node names, JVM heap usage, disk usage, CPU load, and index counts — packaged into an Excel workbook so the proposal deck can reference real numbers.
The architect knows where the data lives. It's in Kibana's Stack Monitoring. The problem is that Stack Monitoring is built for watching, not exporting. Every metric is a chart. There's no "download this table as CSV" button.
The bad version:
- Open Stack Monitoring, click into each node one at a time, read the JVM heap percentage off the gauge, write it in a sticky note.
- Repeat for disk usage. Repeat for CPU. Twelve nodes.
- Switch to Index Management, find the document counts, note them somewhere else.
- Reconcile all the sticky-note data into a single workbook, realize the timestamps don't match because you took 45 minutes to do the manual collection.
By the time you finish, the "point-in-time snapshot" is a 45-minute smear. The proposal needs the numbers from right now.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Excel workbook. It reads Kibana's monitoring APIs and writes the node metrics directly to the workbook — all nodes at the same moment, no manual reading of gauges required.
Fetch Elasticsearch node metrics from Kibana and write each node's name, JVM heap used percentage, disk used percentage, and CPU percent to this sheet, one node per row.
What You Get
- Column A: node name
- Column B: JVM heap used percentage (0–100)
- Column C: disk used percentage (0–100)
- Column D: CPU usage percentage (0–100)
- All 12 nodes captured at the same moment, ready for the proposal
What If the Data Is Not Quite Ready
You also need index storage data for the proposal
Node metrics tell half the story; index sizes tell the other half:
List all indices in Kibana Index Management, write each index's name, status, document count, and storage size to this sheet, and flag any index with more than 50 GB of storage by marking column E as 'Large'.
The proposal needs averages across the cluster, not individual node rows
The VP's deck wants cluster-level summary figures:
From the Elasticsearch node metrics already in this sheet, calculate the average JVM heap used percentage, average disk used percentage, and average CPU percentage across all nodes, and write those three averages as a summary row at the bottom of the data with a 'Cluster Average' label in column A.
You need to flag nodes that are approaching capacity thresholds
The scaling proposal needs to justify which nodes are the bottlenecks:
Fetch Elasticsearch node metrics from Kibana and write each node's name, JVM heap used percentage, disk used percentage, and CPU percent to this sheet. In column E, mark 'At Risk' for any node where JVM heap is above 80% or disk usage is above 75%.
Full capacity snapshot — nodes, indices, and flagged risks in one shot
Fetch Elasticsearch node metrics from Kibana and write each node's name, JVM heap percentage, disk percentage, and CPU percentage to rows 1 through 15 of this sheet. Then fetch all Elasticsearch indices and write their names, document counts, and storage sizes starting at row 17. Flag any node with heap or disk above 80% with 'At Risk' in column E, and any index over 50 GB with 'Large' in column E.
The pattern: capturing the full cluster snapshot — nodes and indices — plus the flagging logic in a single prompt, so the workbook is proposal-ready the moment it runs.
Try It
Get the 7-day free trial of SheetXAI and open any Excel workbook you're building for an infrastructure review or capacity proposal, then ask it to pull a live Elasticsearch cluster snapshot from Kibana. You can also explore Inventory Kibana Fleet Agent Policies Into an Excel workbook or return to the Kibana integration overview.