* Convert ClusterRoles to Roles for all components for multiple namespaces
* Added templates for rules and role generator to dynamically generate roles for each component
* Fix role-generator component templating
* Enable namespaces variable for preupgradechecks
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Co-authored-by: shaunak_deshmukh <shaunak@infracloud.io>
Co-authored-by: Sanket Sudake <sanketsudake@gmail.com>
* Add variables for function/builder namespace in the test
* Allow different namespaces in integration tests
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Co-authored-by: Sanket Sudake <sanketsudake@gmail.com>
Use the Kubernetes and Fission Client from CLI instead of Controller API.
This removes port-forwarding for the controller across Fission CLI mostly.
* Use configurable client in CLI
* Move resource namespace under cmd client
* use server to get fission version
* get archive with URL
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Co-authored-by: Sanket Sudake <sanketsudake@gmail.com>
We allow functions to refer to configmap/secrets. We are monitoring all namespaces for config maps and secrets and also allow cross-namespace references.
This fix monitors configmaps/secret updates in specific namespaces. Also, we ignore cross-namespace references for configmap/secret updates.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* added unit test for setup role binding method
* changes to throw an error in case of mismatch in cluster role name
* changes to add a check for role kind
* Role and RoleBinding for fission-fetcher and fission-builder can be created in additional namespaces
* Added template file for fetcher and builder role and rolebinding
This PR moves fission CLI as closer as possible to kubectl command behaviour. We have improved namespace handling behaviour across CLI.
* add namespace param for fn and env
* use common fn for ns check
* update validation
* default namespace for httpTrigger, env and package, config and triggers
* use default ns
* add namespace filter to spec
* add forceNamespace flag
* set current namespace
* add default namespace in config
* add namespace specific destroy
* add all namespace in the list of resources
* add namespace as global tag
* use %s instead of %v
* add test cases for namespace
* use ns in get all functions
* Separate service accounts for each component
* Permission changes per component
* Fixed fluentbit permissions
* added hooks to pre-upgrade service account and role
* Adjusted hook weights for pre-upgrade hook
* Replaced * with explicit permissions
Added properties to configure object reaper interval, global and specific to exec type.
OBJECT_REAPER_INTERVAL - global
NEWDEPLOY_OBJECT_REAPER_INTERVAL - for new deploy type
CONTAINER_OBJECT_REAPER_INTERVAL - for container type
POOLMGR_OBJECT_REAPER_INTERVAL - for poolmgr
These changes have specifically been made for Kafka connector. This will expose a new metric named `fission_mqt_message_lag`, which will show a number of messages lag per topic and partition.
We can use this metric in the auto-scaling of the pod for the new deploy type executor function. While creating a new deploy function we need to add hpa metrics of external type inside the function definition.
* allow two HTTP triggers with no URL and different prefix
* update dependency
* Fix controller existing tests
* Ensure namespace cleanup in API test
* update test cases
* handle error conditions in test
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Co-authored-by: Sanket Sudake <sanketsudake@gmail.com>
* Add grafana dashboards and configmap creation
* Fix Data source, job and template rules
* Add units to all panels
* fix units and fix linting issues
* Enable verbose linting for dashboards
Co-authored-by: shaunak_deshmukh <shaunak@infracloud.io>
Configmap inside pods for newdeploy and pool manager executor type were not being updated if the user update the configmap.
This fix will help to update the pods for both executor type with new configmap. As per the changes if there is any configmap update then pods will get restarted for both executor type and then it will refer new configmap.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Co-authored-by: Sanket Sudake <sanketsudake@gmail.com>
