- added retainPods flag to take in the number of specialized pods to retain
- add retainPods in both the create function and update function command
- modify crd keys to be typed instead of string
- keep track of function generation in case of update function operation
- add delete handler function to make sure specialized pods are deleted in case function is deleted
---------
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Signed-off-by: Pranoy Kundu <pranoy1998k@gmail.com>
Co-authored-by: Pranoy Kundu <pranoy1998k@gmail.com>
* Add fixes for failure in specialization
* reduce specialization in progress and remove expired requests from queue when specialization is timed out
* rename markSpecializationFailure and remove logger from the queue
* refactor clean up code in api.go and add test case for queue
Details:
- Cleanup svc waiting for the counter in the pool manager if specialization fails
- Cleanup active requests counter in pool manager if client exists the demand for function service while we have allocated function service
- Consider specialization timeout if pod ready timeout > specialization timeout in waiting for ready pod. We also consider if the request to choosePod is cancelled.
- We ensure if we have requests waiting for service requests but if there is no pod in the specialization we clean up those.
---------
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Co-authored-by: Pranoy Kundu <pranoy1998k@gmail.com>
Merged pool cache package into fscache to avoid import cycle.
Also changed all types in pool cache from interface to specific
types.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Use informerfactory across executor
* Run function informer for poolpodcontroller if istio enabled
* Use same namespace for secret as keda mqtriggers
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Remove cluster role references
* Convert secret/package getter cluster roles to role
* Remove all cluster role binding permissions
* Remove unwanted permission
* Try removing all RBAC-related code
* Remove additional constants
* Add functionality for service account checks
* Restrict permission across services
* Remove package informer factory from the executor
* Revert service account check code for now
* Skip adding roles in place of cluster roles
* Remove additional permission from the router
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
We enable security context by default now, so that the fission pod do not have root access to the system.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Backup Prometheus data from CI run
* use `kube-prometheus-stack` instead of `prometheus`
* Add dump-analyzer script
* Fix monitor helm chart values
* Update readme for dump-analyzer
* Disable grafana and alertmanager
Signed-off-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>
* Updated all Go language dependencies to latest version available
* Formatted all files as per gofmt
* Update Golangci-lint version to 1.48.0
* Updated action version wherer application in Github workflows
* Updated Kubernetes version to latest available
* Remove "io/ioutil" references and replace with "io"/"os"
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
1. Use websocket support introduced in Openetelemetry Librarries
2. Use propagators settings introduced in Opentelemetry Go SDK
3. Use samplers settings introduced in Openetelemetry Go SDK
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Add support for custom metrics for HPA
* Cleanup TargetCPUPercent references from possible places
* HPA v2beta has 80% default cpu limit if not set
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Migrate HPA v1 to v2beta2
HPA v2beta2 is defined and supported from 1.19+ onwards.
Also HPA v2 is stable from 1.23 onwards. As we support 1.19+
onwards using HPA v2beta2.
This change is base for custom metrics support we want to add
later by modifying Function spec.
* Add unit tests for hpa operations
* Use constants instead of strings
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Use latest goreleaser with GOAMD64 support
* Update Go version to 1.18.1
* Hardcode GOAMD64 in cli install
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Using interface makes it easy to create a fake client and unit test
a specific portion of the code. We should be able to more write unit
test and increase coverage of code with this change.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
In pool pod controller we were using pool namespace
rather than pod namespace in cleanup which was causing
issue in few scenarios. Using pod namespace now instead.
Also add unit test for scenario which was failing.
Using kubernetes client interface now across instead of
kubernetes ClientSet for testing.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Defining httpserver package to capture httpserver shutdown and
introduces uniform running of http server across codebase.
* Add unit tests for httpserver
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
Currently we dial on localport and wait until, port forwarding
is done. Using channel instead from forwarder to wait explicitly.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
We are removing Fission deprecated Azure Storage Queue connector and
planning to adopt Keda going forward to have better
delegated functionality and more rich support.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
We are removing Fission deprecated Nats connector and
planning to adopt Keda going forward to have better
delegated functionality and more rich support.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
We should avoid tap service call to executor if service URL retrieved from executor is empty.
Added sanity checks to ensure that.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Update Go version to 1.17 and dependencies
* Changes in ProbeHandler and LifeCycleHandler
* Update CRD definitions
* Update Go version to 1.17 in Github Actions
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Removing Prometheus dependency from fission chart
* User should install their own Prometheus installation
* Install Prometheus chart in Github CI workflow
* Change Prometheus query logging to debug
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
This change mainly fixes few things around router and executor
exposed metrices.
1. We are trying to follow standard in metric names.
2. Lables such as namespace are colliding with kube-prometheus standards
so they are getting relabled to exported_namespace. Added function prefix
to resolve this.
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
* Change RBAC resource names work for multiple Fission release
* Fetch secret configmap and package cluster role based on the release name
* Remove default namespace hardcoding from helm chart
Signed-off-by: Sanket Sudake <sanketsudake@gmail.com>
