Selectable Index Patterns for Elasticsearch DataSources

[gelicia]

Discussed in https://github.com/grafana/grafana/discussions/56252

^{Originally posted by hikerspath November 26, 2019} What would you like to be added: Allow elasticsearch index name selection when creating a query for a panel (similar to namespace selection on a CloudWatch datasource.

Why is this needed: Presently, when setting up an elasticsearch datasource, you must enter the index name as part of the datasource configuration. This means that even though accessing the same elasticsearch cluster, I need to manually define every single index pattern as its own datasource.

If instead there was only necessity to define the timeseries pattern format (EX (.*)-YYYY.MM.DD), it would be very easy to provide the user with a selectable list of index patterns that match per query using the GET /_aliases endpoint on an elasticsearch cluster, or perhaps one of the /_cat/ endpoints depending on how the information is parsed.

[cjabrantes]

Hi all,

Can you please update when such feature will be available, this is crucial for me, is the difference of having 1 data source or dozens of them..or not even be able to have the needed information as in the same panel i can only have 1 datasource.

i though this issue was more or less solved with: https://github.com/grafana/grafana/pull/29477

it seems to help a few people according to https://github.com/grafana/grafana/discussions/56252

Besides being able to select the index pattern at query level, its also important to support the remote clusters format: remotecluster-0:my-logs-*,remotecluster-1:my-logs-*, ....

Can you confirm if this is supported at all? i tried to set it in the index pattern and does not seem to work.

So as is i can't have in same panel information from multiple clusters.

Many thanks!

[difeid]

Same here! With this feature we can reduce a lot the number of data sources and implement new use cases if there is remote cluster/index pattern overwrite capabilities in query. It will be nice to have it!

Thanks

[cjabrantes]

Hi,

Is it possible to provide any ETA when this feature will be available and if remote cluster will have be a possibility?

Thanks,

[gelicia]

@cjabrantes It is in the the correct team's backlog, and will be picked up when it is prioritized with their other work. There is no ETA at this time.