Currently all API requests set Cache-control: no-cache to avoid browsers
caching sensitive data. This fixes so that all responses returned from
backend not are cached using http headers. The exception is the data proxy
where we don't add these http headers in case datasource backend needs
to control whether data can be cached or not.
Fixes#16845
* update single stat data model
* update single stat data model
* update single stat data model
* show limit default
* merge master
* change stat selector to single until #15954
* add tooltip
* begin children
* move options under display
* align gauge options
* add migration tests
* Docs: Updated changelog
* SingleStatPanels: show title if manual specified
* FieldPropEditor: Max should change max
* change stats to calcs in config
* remove prefix/suffix
* add test
* abort field cycle when passed the limit
* stub a better test
* move title to Field
* remove title
* try this again
* use element rather than grid position
* adding back console output to debug gridPos alternative
* less logging
* simplify
* subscribe/unsubscribe to event streams when view changes
* Panels: Minor change to lazy loading
A new parameter `queryMode` is added to the InfluxDB datasource to provide a way to use POST instead of GET when querying the database. This prevents to get any error when querying the database with a heavy request.
Default configuration is kept to GET for backward compatibility. Tests and documentation have been added for this new behaviour.
* Added search-query-parser package
* Migrate search input field to react and enable current folter filtering
* Reveiw changes
* FIx tags
* Fix event handlers passed to html elements directly
* noImplicitAny fix
* Debounce search method in search controller
* Search: have clear reset query as well
This fixes a regression introduced by #16724 when datasource is configured
to use HTTP method POST and testing the datasource which results in an
'options.headers is undefined'. Same error probably happens if datasource
is configured to use basic auth and testing datasource.
* speed up circleci build
* convert build.sh to functions and add args for different build types
* add yarn cache
* add build latest to packaging step
* fast build should not try to tag non-amd builds
* remove signing step for prs-and-branches
If Grafana rotates the user's auth token during a request to the data
source proxy it will set the Set-Cookie header with new auth token in
response before proxying the request to the datasource.
Before this fix the Set-Cookie response header was cleared after the
proxied request was finished to make sure that proxied datasources
cannot affect cookies in users browsers. This had the consequence
of accidentally also clearing the new auth token set in Set-Cookie
header.
With this fix the original Set-Cookie value in response header is now
restored after the proxied datasource request is finished. The existing
logic of clearing Set-Cookie response header from proxied request
have been left intact.
Fixes#16757
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
* Add file path to metadata and show it in dialogs
* Make path relative to config directory
* Fix tests
* Add test for the relative path
* Refactor to use path relative to provisioner path
* Change return types
* Rename attribute
* Small fixes from review
* Provisioning: Trying to fix test
* Use better sync strategy for test
* Lower the timeout
* Remove commented code
Co-Authored-By: aocenas <mr.ocenas@gmail.com>
Explore now uses the timezone of the user to decide if local browser
time or UTC should be used.
- Now uses TimeRange instead of RawTimeRange in explore item
state tree and only parsing actual time in a few action
handlers.
- Time picker should now properly handle moving back/forward and
apply time range when both utc and non utc time zone.
- URL range representation is changed from YYYY-MM-DD HH:mm:ss
to epoch ms.
- Now uses AbsoluteTimeRange in graph component instead of moment.
- Makes a copy of the time range passed to timeSrv to make sure immutability
of explore time range when for example elasticsearch test datasources uses
timeSrv and sets a time range of last 1 min.
- Various refactorings and cleanup.
Closes#12812
