Export and import via curl kibana's dashboards and visualizations

Elastic Stack Kibana
1

hi there,

I'm trying to get all the dashboards into json without the UI of the kibana itself (using curl).

at the moment , I succeed getting the dashboards but to open them on another instance I need to have the visualizations inside as well.
My problem is to get the visualizations from kibana and export them to json via curl...
how i can do it? ideas?

one more thing, using the Export Everything on the kibana didn't gave me any json file or anything at all.

Thanks

2 Likes
2

Hi @LiorK,

I am also trying to import dashboard using kibana. Can you tell me how you succeeded as I am trying
curl -XPUT http://localhost:9200/.kibana/dashboard/testdashboard -d '{[
{
"_id": "D1",
"_type": "dashboard",
"_source": {
"title": "D1",
"hits": 0,
"description": "",
"panelsJSON": "[{"id":"V1","type":"visualization","panelIndex":1,"size_x":3,"size_y":2,"col":1,"row":1}]",
"optionsJSON": "{"darkTheme":false}",
"uiStateJSON": "{}",
"version": 1,
"timeRestore": false,
"kibanaSavedObjectMeta": {
"searchSourceJSON": "{"filter":[{"query":{"query_string":{"query":"*","analyze_wildcard":true}}}]}"
}
}
}
]}'
and getting error as :
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 1024 100 469 100 555 14656 17343 --:--:-- --:--:-- --:--:-- 17343{"error":{"root_cause":[{"type":"mapper_parsing_exception","reason":"failed to parse"}],"type":"mapper_parsing_exception","reason":"failed to parse","caused_by":{"type":"json_parse_exception","reason":"Unexpected character ('[' (code 91)): was expecting either valid name character (for unquoted name) or double-quote (for quoted) to start field name\n at [Source: org.elasticsearch.common.io.stream.InputStreamStreamInput@77224b66; line: 1, column: 3]"}},"status":400}

Kindly suggest some points.

3

I use a Python script that pulls the dashboards I want from ES, enumerates the panels in the JSON of the dashboard, grabs those visualisations from ES and then looks for the index-patterns each of the visualisations uses and grabs those. Once I have all these documents I index them to the target index/ES host.

Right now I'm using an index template to avoid dynamic mapping issues (which I have seen), in the future I'll be pulling the mapping from the source index which will be more robust in supporting new Kibana features and plugins.

4

currectly i have script importing all the dashboards, visualisations and searches. and exporting to another .kibana index.

@Rajeev_Gaurav

  • Note- gettin the dashboards alon will not help you if it has visualisations and also visualisations that uses searches. you will need those objects as well.
  • your dashboard D1 hav visualisations inside? can you try to reach the following url http://localhost:9200/.kibana/dashboard/D1/_source

@pemontto can you please explain how dynamic mapping is related to this?

5

Assuming you understand Elasticsearch dynamic mapping, an examples where this goes awry is where Kibana has a time field, specifically the "timeFrom " and "timeTo" in the dashboard document type. These fields can be a full ISO8601 date or relative like "now-3y".

If Elasticsearch first sees a document with an ISO8601 date the field will be dynamically set to a date field, now when you attempt to index that document with the "now-3y" you'll get a data type mismatch and it will fail to index all those dashboards.

So the solution is to either copy across the mapping from the source index, or use an index-template. Again I'd suggest updating the mapping from the source for future proofing.

6

I tried https://github.com/elastic/beats-dashboards but problem with this is that it does not support source meta data ("_id") and thus It is not usable as i was using meta data for uniqueness of visualization and their mapping in dashboard.So can suggestv points on that.

7

I would post this question in the Beats channel: https://discuss.elastic.co/c/beats

8

How to export and import he pre configured dashboards Kibana ( 5.4.0)
can pls anyone help me