Adds a config section with derived fields which is a config that allows you to create a new field based on a regex matcher run on a log message create DataLink to it which is the clickable in the log detail.
* POC: trying to see if there is a way to support objects in template interpolations
* Added support for nested objects, and arrays
* Added accessor cache
* fixed unit tests
* First take
* Use links supplier in graph
* Add field's index to cache items
* Get field index from field cache
* CHange FiledCacheItem to FieldWithIndex
* Add refId to TimeSeries class
* Make field link supplier work with _series, _field and _value vars
* use field link supplier in graph
* Fix yaxis settings
* Update dashboard schema version and add migration for data links variables
* Update snapshots
* Update build in data link variables
* FieldCache - idx -> index
* Add current query results to panel editor
* WIP Updated data links dropdown to display new variables
* Fix build
* Update variables syntac in field display, update migration
* Field links supplier: review updates
* Add data frame view and field name to TimeSeries for later inspection
* Retrieve data frame from TimeSeries when clicking on plot graph
* Use data frame's index instead of view
* Retrieve data frame by index instead of view on TimeSeries
* Update data links prism regex
* Fix typecheck
* Add value variables to suggestions list
* UI update
* Rename field to config in DisplayProcessorOptions
* Proces single value of a field instead of entire data frame
* Updated font size from 10px to 12px for auto complete
* Replace fieldName with fieldIndex in TimeSeries
* Don't use .entries() for iterating in field cache
* Don't use FieldCache when retrieving field for datalinks in graph
* Add value calculation variable to data links (#19031)
* Add support for labels with dots in the name (#19033)
* Docs update
* Use field name instead of removed series.fieldName
* Add test dashboard
* Typos fix
* Make visualization tab subscribe to query results
* Added tags to dashboard so it shows up in lists
* minor docs fix
* Update singlestat-ish variables suggestions to contain series variables
* Decrease suggestions update debounce
* Enable whitespace characters(new line, space) in links and strip them when processing the data link
* minor data links UI update
* DataLinks: Add __from and __to variables suggestions to data links (#19093)
* Add from and to variables suggestions to data links
* Update docs
* UI update and added info text
* Change ESC global bind to bind (doesn't capture ESC on input)
* Close datalinks suggestions on ESC
* Remove unnecessary fragment
* WIP: Use data frames in explore
* Explore: everything seems to be working again
* Reworked ResultProcessor tests
* Fixed unit test
* Add some typings and comments
* add appending utility
* add appending utility
* update comment
* rename to mutable
* move mutable functions out of DataFrameHelper
* move mutable functions out of DataFrameHelper
* move mutable functions out of DataFrameHelper
* turn DataFrameHelper into FieldCache
* guess time from name
* graph the numbers
* return the timeField, not just the index
* just warn on duplicate field names
* only use a parser if the input is a string
* append init all fields to the same length
* typo
* only parse string if value is a string
* DataFrame: test fixes
* Switch to null for missing values
* Fixed tests
* Wip: Compiles and runs
* WIP: Logs Graph partially working
* Refactor: Adds GraphSeriesToggler
* Refactor: Adds tickDecimals to YAxis
* Refactor: Adds TimeZone and PlotSelection to Graph
* Refactor: Makes the graphResult work in Explore
* Refactor: Adds ExploreGraphPanel that is used by Logs and Explore
* Fix: Fixes strange behaviour with ExploreMode not beeing changed
* Fix: Adds onSelectionChanged to GraphWithLegend
* Refactor: Cleans up unused comments
* ExploreGraph: Disable colorpicker
* Refactor: Removes switches for TimeStamp and LocalTime and displays time as configured for user
* Styles: Updates classname used to display timeUtc
* Refactor: Adds switch for Time column
If a field or a label named level is returned from datasource that is
used as log level for the logs result instead of parsing the log level
from the message.
Closes#17122
* Move log's typings into grafana/ui
* Update the way context is retrieved for Loki
Major changes:
1. getLogRowContext expects row to be of LogRowModel type
2. getLogRowContext accepts generic options object, specific to a datasource of interest. limit option has been removed, and now it's a part of Loki's context query options (see below)
3. LogRowContextProvider performs two queries now. Before, it was Loki ds that performed queries in both directions when getLogRowContext.
4. Loki's getLogRowContext accepts options object of a type:
interface LokiContextQueryOptions {
direction?: 'BACKWARD' | 'FORWARD';
limit?: number;
}
This will enable querying in either direction independently and also slightly simplifies the way query errors are handled.
LogRowContextProvider maps the results to a Loki specific context types, basically string[][], as raw log lines are displayed in first version.
* (feat/explore): Support for new LogQL filtering syntax
Loki is adding syntax to support chained filtering.
This PR adapts Grafana to support this.
- Send only `query` parameter in loki request
- Automatically wrap search text in simple syntax, e.g., `{} foo` is
sent as `{} |~ "foo"`.
* Adapted to regexp parameter staying on in Loki
* Dont wrap single regexp in new filter syntax
* Fix datasource test
* Fallback regexp parameter for legacy queries
* Fix search highlighting
* Make highlighting work for filter chains
* Fix datasource test
This is the first step moving towards Explore supporting logs for
more datasources than Loki. In the first step we move all the log
processing from Loki into Explore.
- Make explore convert logs result returned from datasources to SeriesData,
if needed, and for now convert SeriesData to LogsModel.
- Loki datasource query now returns SeriesData and all
processing have been moved into explore instead.
- Removed key from LogRowModel and use log row indexes as
the unique key instead.
- Removed id from LogsModel since it looks like it's not in use.
- Introduced a new class FieldCache which is convenient to use when
looking up multiple fields and/or field types and series values.
Closes#16287
The log rows were mutated in the store, what resulted in wrong deduplication results as well as made the selector memoization quirky - for "the same results" the selector was running twice.
- changed log parser API to be more semantic: getting fields, getting values
- regexps become an implementation detail of parser and are no longer part of the API
- improved JSON parser to not be tripped by nested quotation marks
- improved JSON parser to support numbers
- added more tests
Lazily parse lines and display stats for field when clicked on.
- log line parsers for JSON (basic), and logfmt
- delayed parsing in case user moves to other line
- reuse label stats for field stats
- fixed typo
- moved result calculation to explore utils
- use component keys instead of componentWillReceiveProps
- require logs to have and id, to as use as Logs component key
- render delay based on row count
