Look for 'caused_by.reason' in ES error response
When the ES response does not contain `reason` or `root_cause[0].reason`
is empty, there is no information for the user to know what is going
wrong.
An example of the error message after this change:
```
Failed to evaluate queries and expressions: failed to execute query A: Trying to create too many buckets. Must be less than or equal to: [65536] but this number of buckets was exceeded. This limit can be set by changing the [search.max_buckets] cluster level setting.
```
Related to https://github.com/grafana/grafana/issues/61246
* Return errors from data parsing
* Better error handling
* Fix the tests
* When there is no frame add empty frame to get metadata attached to it
* Fix tests
* Update testdata
* use new log group picker also for non cross-account queries
* cleanup and add comment
* remove not used code
* remove not used test
* add error message when trying to set log groups before saving
* fix bugs from pr feedback
* add more tests
* fix broken test
* Correctly set filter values in portal URL
* Refactor to include dimensions as a part of AzureMonitor query
* Correctly set splitting value in URL
- Add type for dimension filters object
* Update tests
* Don't test dimensions
Automatically forward core plugin request HTTP headers in outgoing HTTP requests.
Core datasource plugin authors don't have to specifically handle forwarding of HTTP
headers, e.g. do not have to "hardcode" the header-names in the datasource plugin,
if not having custom needs.
Fixes#57065
* GoogleCloudMonitoring: Migrate queries to the new format
* Refactor Aligment and AligmentFunction components (#60235)
* Adapt CloudMonitoringDatasource and CloudMonitoringAnnotationSupport (#60177)
* Fix: avoid migration for new queries (#60375)
* Move preprocessor handling to the backend (#60383)
* Other fixes and new function (#60411)
* Adapt components to the new API (#60451)
* Split metrics query type in time series list and query (#60475)
* Clean up metricQuery references (#60478)
* More bug fixes (#60525)
* SQL Datasources: Use health check for config test
* Remove unnecessary test
* Fix test errors
* Revert mysql go driver update
* Use transform query error
* Use TransformQueryError from sql_engine
* Refactor parse query to functions
* Move parsing to new file
* Create empty result variable and use it when returning early
* Fix linting
* Revert "Create empty result variable and use it when returning early"
This reverts commit 36a503f66e.
* Elasticsearch: Fix ordering in raw_document and add logic for raw_data
* Add comments
* Fix raw data request to use correct timefield
* Fix linting
* Add raw data as metric type
* Fix linting
* Elasticsearch: Add defaults for log query
* Add higlight
* Fix lint
* Add snapshot test
* Implement correct query for logs
* Update
* Adjust naming and comments
* Fix lint
* Remove ifs
* Elasticsearch: Fix ordering in raw_document and add logic for raw_data
* Add comments
* Fix raw data request to use correct timefield
* Fix linting
* Add raw data as metric type
* Fix linting
* Hopefully fix lint
* Datasource settings: Add deprecation notice for database field
* SQL Datasources: Migrate from settings.database to settings.jsonData.database
* Check jsonData first
* Remove comment from docs
