"Saved field parameter is now invalid, please select a new field"

Hello,

When we tried to run the Dashboard with Broadsoft, dashboard collaborate and the desktop sharing dashboards, we get the error saying "Saved field parameter is now invalid, please select a new field", we are using the last version of broadsoft scripts and 5.2.2 Kibana and elastics.

Can you help me, please.

Hello @Tania_Bonilla,

I am not familiar with Broadsoft, but I assume that it allowed you to import existing dashboards and visualizations. The current version of Kibana relies on the field_stats api exposed by Elasticsearch. That dependency creates an unfortunate bug where for fields in indices that don't have any data, Kibana recognizes them as unsearchable and not aggregatable. This in turn causes any visualizations that are dependent on those fields to throw the error you mentioned above.

There is an open issue that you can follow for the more comprehensive solution to this problem.

For a more short-term fix, you can 'seed' your new index with a sample document that contains data for each field, and that should allow you to load the visualizations/dashboards without this error.

1 Like

Thank you for your response but, How can I do " For a more short-term fix, you can 'seed' your new index with a sample document that contains data for each field, and that should allow you to load the visualizations/dashboards without this error."

My Index

Does the index that this visualization points to contain data?

When you go to the Index Patterns section of Kibana, and select that index pattern, does the field that the visualization depends on show as searchable and aggregatable?

I assume that the answer to both of these questions is 'no', so we go back to my suggestion to seed the index. What I mean is, just make sure there is at least one document indexed into the index. The nature of your index/data will determine how you would accomplish that. You could go into Dev Tools and manually index a document, or make sure that however you are populating your index has processed and indexed some documents before attempting to view the visualization in Kibana.

Thank you for you response, but I do not understand, how can I do for upload manually my document. My index

In my Dev Tools

Can you help me please.

Thank you

Can we limit the scope of your issue a little bit? Can you find a single visualization and view it in the visualization editor to reproduce the error?

The error appears in

and in

Thank you!

What is the result if you run this query in Dev Tools?

POST .kibana/visualization/_search
{
  "query": {
    "match": {
      "title": "Room Type"
    }
  }
}

Thank you ! The result is :

{
"took": 126,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"failed": 0
},
"hits": {
"total": 2,
"max_score": 7.2276917,
"hits": [
{
"_index": ".kibana",
"_type": "visualization",
"_id": "Col:-Room-Type",
"_score": 7.2276917,
"_source": {
"title": "Room Type",
"visState": """
{
"title": "Room Type",
"type": "pie",
"params": {
"addLegend": true,
"addTooltip": true,
"isDonut": false,
"shareYAxis": true
},
"aggs": [
{
"id": "1",
"type": "count",
"schema": "metric",
"params": {}
},
{
"id": "2",
"type": "terms",
"schema": "segment",
"params": {
"field": "collaborate_roomtype",
"size": 10,
"order": "desc",
"orderBy": "1"
}
}
],
"listeners": {}
}
""",
"uiStateJSON": "{}",
"description": "",
"version": 1,
"kibanaSavedObjectMeta": {
"searchSourceJSON": """
{
"index": "bwcdr*",
"query": {
"query_string": {
"analyze_wildcard": true,
"query": "collaborate_role:"
}
},
"filter": []
}
"""
}
}
},
{
"_index": ".kibana",
"_type": "visualization",
"_id": "Sharing:-room-joins-per-time-interval",
"_score": 2.0183094,
"_source": {
"title": "Sharing: room joins per time interval",
"visState": """{"aggs":[{"id":"2","params":{"customInterval":"2h","extended_bounds":{},"field":"logtimestamp","interval":"auto","min_doc_count":1},"schema":"segment","type":"date_histogram"},{"id":"3","params":{"customLabel":"Joins"},"schema":"metric","type":"count"},{"id":"4","params":{"filters":[{"input":{"query":{"query_string":{"analyze_wildcard":true,"query":"source: participant"}}},"label":"Participants"},{"input":{"query":{"query_string":{"analyze_wildcard":true,"query":"source: owner"}}},"label":"Owners"}]},"schema":"group","type":"filters"}],"listeners":{},"params":{"addLegend":true,"addTimeMarker":false,"addTooltip":true,"defaultYExtents":false,"mode":"stacked","scale":"linear","setYExtents":false,"shareYAxis":true,"times":[],"yAxis":{}},"title":"Sharing: room joins per time interval","type":"histogram"}""",
"uiStateJSON": """{"vis":{"legendOpen":true}}""",
"description": "",
"version": 1,
"kibanaSavedObjectMeta": {
"searchSourceJSON": """{"index":"bwlog
","query":{"query_string":{"query":"application:USSLog AND channel: RoomControl AND ussroomid:* AND source:"successfully joined room"","analyze_wildcard":true}},"filter":[]}"""
}
}
}
]
}
}

Next:

  1. Go to Management/Index Patterns.
  2. Select the bwlog* index pattern
  3. In the 'Filter' box of the field list type 'collaborate_roomtype'
  4. What do the searchable and aggregatable columns show for that field? (check or no check)

No check

Execute this is Dev Tools:

POST bwlog_seed/seed/seed_doc
{
  "collaborate_roomtype": "test"
}

Then:

  1. Go to Management/Index Patterns.
  2. Select the bwlog* index pattern
  3. Click the Refresh Field List button.
  4. What do the searchable and aggregatable columns show for the collaborate_roomtype field now? (check or no check)

Yes now, are check

Try to reload the 'Room Type' visualization.

Now, no errors in the room type, but how I change all the others Visualize.

Thank you!

To summarize what I did.

  1. Find the visualization that is throwing the error
  2. Query for the definition of that visualization in Dev Tools
  3. Look at the searchSourceJSON property to determine the index pattern that was used in the visualization
  4. Look at the visState property to find which fields were used in the aggregations.
  5. Make sure the field is the one causing the problem by looking at whether those checkboxes appear in the field list for that field
  6. If the checkboxes are not there, index a throwaway document into the respective index, and refresh the fields.
  7. Rinse, repeat
2 Likes

One Post and Query , for each.?
Thank you!

If they are using the same index pattern, you can index all of the fields in one document. The real issue is that there needs to be at least one document with the field populated in your index for each field that is referenced by your visualizations.

I create the POST
POST .kibana/visualization/_search
{
"query": {
"match": {
"title": "unique owners"
}
}
}
POST bwcdr_seed/seed/seed_doc
{
"collaborate_uniquer-owners ": "test"
}
but the result is

Did you click the 'Refresh Field List' button in Management/Index Patterns?