Kibana Saved Objects API Skill#

Manage Kibana Saved Objects including dashboards, Lens visualizations, and other Kibana entities via REST API.

Base URL#

text

http://<kibana-host>/api/saved_objects

For Omni-Monitor: http://192.168.99.43/api/saved_objects

Required Header: kbn-xsrf: true

---

Core Endpoints#

1. Create Saved Object#

text

POST /api/saved_objects/<type>

Body:

json

{
  "attributes": {
    "title": "My Dashboard",
    "description": "..."
  },
  "id": "optional-custom-id",
  "references": [{"id": "...", "name": "...", "type": "index-pattern"}]
}

2. Get Saved Object#

text

GET /api/saved_objects/<type>/<id>

3. Update Saved Object#

text

PUT /api/saved_objects/<type>/<id>

Body:

json

{
  "attributes": {
    "title": "Updated Title"
  },
  "references": [...]
}

4. Delete Saved Object#

text

DELETE /api/saved_objects/<type>/<id>

5. Bulk Get#

text

POST /api/saved_objects/_bulk_get

Body:

json

{
  "objects": [
    {"type": "dashboard", "id": "abc123"},
    {"type": "lens", "id": "def456"}
  ]
}

6. Bulk Create#

text

POST /api/saved_objects/_bulk_create

Body:

json

{
  "objects": [
    {"type": "dashboard", "id": "...", "attributes": {...}},
    {"type": "lens", "id": "...", "attributes": {...}}
  ]
}

7. Export Saved Objects#

text

POST /api/saved_objects/_export

Body:

json

{
  "objects": [
    {"type": "dashboard", "id": "abc123"},
    {"type": "lens", "id": "def456"}
  ]
}

Returns: NDJSON stream of saved objects

8. Import Saved Objects#

text

POST /api/saved_objects/_import?overwrite=true

Content-Type: multipart/form-data Form field: file = NDJSON file

9. Find (Search) Saved Objects#

text

POST /api/saved_objects/_find

Body:

json

{
  "type": "dashboard",
  "search": "title keyword",
  "searchFields": ["title", "description"],
  "page": 1,
  "perPage": 20
}

---

Common Object Types#

TypeDescriptiondashboardKibana dashboardslensLens visualizationsvisualizationClassic visualizations (TSVB, etc.)index-patternData views (index patterns)searchSaved searchesmapMap visualizationstagTags for organizingcanvas-workpadCanvas workpads

---

References System#

References link objects together. For example, a dashboard panel (lens) references:

1. The index-pattern (data view) it uses
2. Any parent dashboard it belongs to

Reference format:

json

{
  "id": "b58d25c5-c05c-47be-a6cb-4073ef478a8f",
  "name": "indexpattern-datasource-layer-aca264a7-bf68-47b3-af63-db1e67fe0646",
  "type": "index-pattern"
}

Creating with references:

json

{
  "type": "lens",
  "attributes": {
    "title": "CPU Usage",
    "state": {
      "datasourceStates": {...},
      "visualization": {...}
    }
  },
  "references": [
    {"id": "b58d25c5-c05c-47be-a6cb-4073ef478a8f", "name": "indexpattern-datasource-layer-xxx", "type": "index-pattern"}
  ]
}

---

Dashboard Structure#

Dashboard attributes:

json

{
  "title": "Dashboard Name",
  "description": "Optional description",
  "panelsJSON": "[{\"type\":\"lens\",\"gridData\":{...},\"embeddableConfig\":{...}}]",
  "optionsJSON": "{\"useMargins\":true,\"syncColors\":false,\"hidePanelTitles\":false}",
  "timeRestore": true,
  "timeTo": "now",
  "timeFrom": "now-24h",
  "refreshInterval": {
    "pause": false,
    "value": 300000
  }
}

---

Lens Object Structure#

Lens is Kibana's next-generation visualization. Key structure:

json

{
  "type": "lens",
  "attributes": {
    "title": "Metric Title",
    "state": {
      "datasourceStates": {
        "formBased": {
          "layers": {
            "<layer-uuid>": {
              "columns": {
                "<accessor-uuid>": {
                  "label": "Metric Label",
                  "dataType": "number",
                  "operationType": "count|average|sum|max|min",
                  "sourceField": "field.name",
                  "isBucketed": false,
                  "scale": "ratio"
                }
              },
              "columnOrder": ["<accessor-uuid>"],
              "incompleteColumns": {}
            }
          }
        }
      },
      "visualization": {
        "layerId": "<layer-uuid>",
        "layerType": "data",
        "metricAccessor": "<accessor-uuid>"
      }
    }
  },
  "references": [
    {"id": "<index-pattern-id>", "name": "indexpattern-datasource-layer-<layer-uuid>", "type": "index-pattern"}
  ]
}

lnsMetric (single metric):

json

{
  "layerId": "...",
  "layerType": "data",
  "metricAccessor": "<accessor-uuid>"
}

lnsXY (line/bar/area chart):

json

{
  "layerId": "...",
  "layerType": "data",
  "layers": [{
    "layerId": "...",
    "accessors": ["<accessor-uuid>"],
    "position": "top",
    "seriesType": "bar|line|area",
    "showGridlines": false,
    "xAccessor": "<accessor-uuid>"
  }]
}

---

Important Rules#

1. Never use __records___ as sourceField → causes "Field not found" errors
2. Text fields must use .keyword suffix for aggregations (e.g., host.keyword)
3. References must use correct format → indexpattern-datasource-layer-<layer-uuid>
4. Lens layerId and accessor UUIDs must be unique per visualization

---

NDJSON Import/Export Format#

Export produces:

text

{"attributes":{"title":"Dash 1"...},"id":"abc","type":"dashboard","references":[...]}
{"attributes":{"title":"Lens 1"...},"id":"def","type":"lens","references":[...]}

Import: Same format, overwrite=true replaces existing objects.

---

Scripts#

ScriptPurposelist_dashboards.pyList all dashboardsexport_dashboard.pyExport a dashboard and its dependenciesimport_ndjson.pyImport NDJSON file to Kibanacreate_lens_metric.pyCreate a simple metric Lens visualization

---

See Also#

• kibana-data-views - Manage data views (index patterns)
• kibana-lens-builder - Generate Lens visualizations for monitoring

---

Troubleshooting#

"Cannot read properties of undefined (reading 'indexpattern')"#

Cause: Dashboard/lens references a non-existent data view.

Fix:

1. List data views: GET /api/data_views
2. Verify referenced data view ID exists
3. Recreate missing data views with kibana-data-views

"Field not found" errors#

Cause: Using text field without .keyword suffix.

Fix: Use host.keyword not host, metric_type.keyword not metric_type

Corrupted saved objects (404 on GET but appears in list)#

Fix:

1. Delete the corrupted object: DELETE /api/saved_objects/<type>/<id>
2. Recreate with proper data