* Implement initial check with schema for generic resources
* Implement List and add tests
* Add namespace type and change to folder_resource name
* Handle namespace grants for typed resources
* Run tests as integration tests
* Add support for verb in list requests
* FIX: Remove the checks for lbac rules inside of datasources
* Remove json validation for lbac rules
* Preserve lbac rules in updates
* Refactored test to remove the table structure
* refactor: change to allow naming and concise override instead of complex branching
* refactor to make sure we set an empty field for updates
* bugfix
* check for datasources.JsonData
* fix merge
* add datasource to check for field presence only
* add function call for readability
* Introduce new models RoutingTree, RouteDefaults and Route and api-server to serve them that is backed by provisioning notification policy service.
* update method UpdatePolicyTree of notification policy service to return route and new version
* declare new actions alert.notifications.routes:read and alert.notifications.routes:write and two corresponding fixed roles.
---------
Co-authored-by: Tom Ratcliffe <tom.ratcliffe@grafana.com>
Co-authored-by: Matthew Jacobson <matthew.jacobson@grafana.com>
* extracted in-proc mode to #93124
* allow insecure conns in dev mode + refactoring
* removed ModeCloud, relying on ModeGrpc and stackID instead to discover if we're running in Cloud
* remove the NamespaceAuthorizer would fail in legacy mode. It will be added back in the future.
* use FlagAppPlatformGrpcClientAuth to enable new behavior, instead of legacy
* extracted authz package changes in #95120
* extracted server side changes in #95086
---------
Co-authored-by: Gabriel MABILLE <gamab@users.noreply.github.com>
Co-authored-by: gamab <gabriel.mabille@grafana.com>
Co-authored-by: Dan Cech <dcech@grafana.com>
* fix: Change users permissions search to use a consistent key without collisions
* Move HashString to cacheutils
* Change error handling logic for what to do with a cache key
* Add a test that confirms search cache key consistency
* add admin permissions upon creation of a folder w. SA
* Update pkg/services/folder/folderimpl/folder.go
Co-authored-by: Karl Persson <kalle.persson@grafana.com>
* Grant service account permissions for creation of dashboards
* Grant service account admin permissions upon creating a datasource
* fetch user using the userservice with the userid
* Revert "fetch user using the userservice with the userid"
This reverts commit 23cba78752.
* revert back to original datasource creation
---------
Co-authored-by: Karl Persson <kalle.persson@grafana.com>
Previously all receiver modifications were denied with alertingApiServer
enabled. This allows pure creates and deletes through as these specific
cases can be handled simply and without risk of rbac shenanigans.
* Fix: Fix panic when json data are nil
* Use Interface()
* Feedback
Co-authored-by: Eric Leijonmarck <eric.leijonmarck@gmail.com>
* Need to check inside the if statement
---------
Co-authored-by: Eric Leijonmarck <eric.leijonmarck@gmail.com>
* Implement uidToResourceID
* add middleware
* Move uidToResourceID to alerting package
* Only hash uid if it's too long
* Use hashed uid in access control
* Move ReceiverUidToResourceId to ScopeProvider
* resolve uid in middleware only if param exists
* Tests
* Linting
---------
Co-authored-by: Yuri Tseretyan <yuriy.tseretyan@grafana.com>
* Remove SettingProvider settings from SSO interactions
* Mock Settings Provider for SSO Settings test
* Ignore error from SettingsProvider
* Add test for backend
* start on tokens
* more error messages
* more handling
* rephrased with suggestions from Daniel
* separate gms parse method
* use translation
* refactor initial idea to use error obj
* use error dto result
* handle gms client
* clean logs and comments
* fix tests
* tests for gms
* test and lint
* lint
* one more handling from gms
* typing in fe
* use error interface
* use validation error
* remove unused gms error
* use errorlib and helper function in fe
* regen api
* use same error util
* one more error to handle
* Also validate folder on provisioning update
* Move folder check before auth check
When checking for the existence of a folder we go through the folder
service which requires auth. Doing so prevents an unprivileged user from
accessing information about whether a particular folder exists or not.
* Rewrite zanzana collector to fetch all available pages
* Register access control as a background service
* If zanzana is enabled we run Syncs and start Reconciliation job
* Update pkg/services/authz/zanzana/client/client.go
Co-authored-by: Alexander Zobnin <alexanderzobnin@gmail.com>
* Use server lock when doing performing reconciliation
* start on loading the error code
* error code to message mapping
* use resource code type
* use defined error code
* partial updates from comments
* i18nKey gen
* fixed t
* fixed translations
* typing
* CloudMigrations: create snapshot for Notification Policy
* CloudMigrations: add notification policy constants and components
* CloudMigrations: add uid to resources that have it
