Fix unexpected error when creating a new cloudwatch datasource.
Involves a fair amount of refactoring, so if this causes unexpected issues related to region fetching we can turn this off with the cloudwatchNewRegionsHandler feature toggle, although we do not predict it will so we are enabling it to default to true and hope to remove it shortly.
* issue loki queries in parallel so total query time is only the slowest query rather than the sum of all query times.
* Fix lint
* Add running of queries in parallel behind feature toggle to test the functonality before release
* Add span end
* Move shared logic to separate function
* Add logging and tracing around running of all queries
---------
Co-authored-by: Ivana Huckova <ivana.huckova@gmail.com>
* Elasticsearch: Add tracing do data source
* Fix tests
* Address feedback
* Update pkg/tsdb/elasticsearch/response_parser.go
Co-authored-by: Sven Grossmann <sven.grossmann@grafana.com>
* Update pkg/tsdb/elasticsearch/response_parser.go
Co-authored-by: Sven Grossmann <sven.grossmann@grafana.com>
* Track error across both spans
* Add span for decoding of response
* Fix test
* Update setting of errors + fix test
---------
Co-authored-by: Sven Grossmann <sven.grossmann@grafana.com>
* Handle the response with different field key order
* More unit tests to cover edge cases
* Cover more edge cases
* make it simpler
* Better test inputs
* Elasticsearch: Adjust naming in logging according to convention
* Log response parsing per response
* Update
* Fix logging of errors when no response
* Add path to error loggigng
* Update pkg/tsdb/elasticsearch/response_parser.go
* adjust Loki to logging convention
* Fix call resource logging
* Update dataquery
* Update
* Remove redundant logging
* Fix TODO
* Rename action to stage and use variables
* `resp` might be `nil`
* `resp` might be `nil` here as well
* change to `statusCode`
* use correct logger
* also here
* add query information to logging
---------
Co-authored-by: Ivana Huckova <ivana.huckova@gmail.com>
* Add support for MI authentication to MSSQL
This adds support for managed identity authentication for MSSQL managed
instances running in Azure.
Co-authored-by: baldm0mma <jev.forsberg@grafana.com>
* removed infra logs
* improve health check
* remove debug and error logs
* feedback
* Update pkg/tsdb/azuremonitor/azuremonitor-resource-handler.go
* Update pkg/tsdb/azuremonitor/loganalytics/azure-log-analytics-datasource.go
* fix close body error
* update test
* resource request should return errors
* go linter
* go linter
* Unify default value
* Use variable to keep default precisions in sync
* Use default precision variable
* Update precision description
* Update defaultPrecisionString and move
* Be more specific in naming of variabkle
* Revert "Merge remote-tracking branch 'origin' into ivana/es-precision-default-value"
This reverts commit 599f236a77, reversing
changes made to 6742be0c6d.
* Revert wrong merge
* Revert wrong merge with turned off lefthook
* Metrics summary
* Update query
* Remove colors
* Update states
* Add group by into its own component
* Add group by to search and traceql tabs
* Add spacing for group by
* Update span kind values
* Update span status code values
* Update query based on target + group by
* Cleanup
* Only add targetQuery if not empty
* Add kind=server to table
* Update groupBy query logic
* Add feature toggle
* Use feature toggle
* Self review
* Update target query
* Make gen-cue
* Tweak query
* Update states
* useRef for onChange
* Fix for steaming in search tab
* Add loading state tests
* metricsSummary tests
* Datasource tests
* Review updates
* Update aria-label
* Update test
* Simplify response state
* More manual testing and feedback from sync call
* Prettier and fix test
* Remove group by component from traceql tab
* Cleanup, tests, error messages
* Add feature tracking
- The util/converter Prometheus response json parse was not checking for errors while parsing. It now does. In particular, if `[dataproxy]/response_limit` is set in Grafana's config, it will now recognize the limit error.
- Fixes#73747
- Adds `jsonitere` package, which wraps json-iterator/go's Iterator's Methods with methods that return errors, so errcheck linting can be relied upon
- Impact:
- If something was sending malformed JSON to the prometheus or loki datasources, the previous code might have accepted that and partially processed the data
- Before there may have been partial data with no error, where as no there may be errors but they will have no partial results, just the error.
