KEDA: Kubernetes-based Event Driven Autoscaling

Releases: You can find the latest official releases and download links at the KEDA Releases page.

Roadmap: The project uses GitHub issues to maintain its backlog. A complete overview of all open items and planning can be found in the ROADMAP.md file.

Sources: README.md

Join the Conversation:

• Slack: Join the #KEDA channel on the Kubernetes Slack to learn or chat about KEDA.
• Meetings: Participate in community meetings to contribute or discuss the project's direction. Details are available on the KEDA website.

Support Policy: For details on the KEDA support policy, visit keda.sh/support/.

Adopters: If you are running KEDA in production, you can become a listed user by following the instructions at kedacore/keda-docs#become-a-listed-keda-user.

Governance: Learn about the project's governance structure at kedacore/governance.

Sources: README.md

Retrieve the URL of the gRPC server that this client is currently connected to.

Usage:

import (
    "github.com/kedacore/keda/v2/pkg/metricsservice"
)

serverURL := client.GetServerURL()
fmt.Printf("Connected to: %s\n", serverURL)

Sources: pkg/metricsservice/client.go

• Complete the setup for Custom KEDA locally outside cluster (generate certs, deploy CRDs).
• Expose your local instance to the internet (e.g., using localtunnel):

  lt --port 9443 --local-https --allow-invalid-cert

Block until the gRPC connection to the KEDA Metrics Server is established and ready. This is useful during initialization to ensure the client can communicate before proceeding.

Usage:

import (
    "context"
    "github.com/go-logr/logr"
    "github.com/kedacore/keda/v2/pkg/metricsservice"
)

// logger: logr.Logger instance
ready := client.WaitForConnectionReady(ctx, logger)
if !ready {
    // Context timeout or connection failed
    return fmt.Errorf("connection not ready")
}

Sources: pkg/metricsservice/client.go

Subscribe to a raw metric stream for a specific metric. This registers the subscriber with the KEDA Metrics Server so that metric updates are pushed to the client.

Prerequisites:

• The client must be initialized with rawStream=true in NewGrpcClient.
• You must call GetRawMetricsStream to open the connection before subscribing.

Usage:

import (
    "context"
    "github.com/kedacore/keda/v2/pkg/metricsservice"
)

// subscriber: Unique identifier for this subscriber
// scaledObjectName: Name of the ScaledObject
// scaledObjectNamespace: Namespace of the ScaledObject
// metricName: Name of the metric to subscribe to

ack, err := client.Subscribe(
    ctx,
    "my-subscriber-id",
    "my-scaled-object",
    "default",
    "my-metric",
)
if err != nil {
    // handle error
}

if ack {
    // Subscription successful
}

Sources: pkg/metricsservice/client.go

To enable memory profiling for KEDA v2, refer to the specific guide provided by the community:

• Guide: Enabling Memory Profiling on KEDA v2

This resource details the steps required to activate profiling features for debugging memory usage.

Sources: BUILD.md

1. Clone the repository:

   git clone git@github.com:kedacore/keda.git
   cd keda
   code .

2. In VS Code, run CTRL+SHIFT+P and select Dev Containers: Reopen in container.
3. Once the container is ready, use the integrated terminal to build:

   make build

Note: The first build may take time to install tooling. The image is cached for subsequent builds.

git clone git@github.com:kedacore/keda.git
cd keda
code .
# In VS Code: CTRL+SHIFT+P -> Dev Containers: Reopen in container
make build

Sources: BUILD.md