Shipped · DevOps and infrastructure

What kedacore/keda shipped

The public repository of Keda · github.com/kedacore/keda

Written by FoxPlug from the repository's own releases; not affiliated with Keda. Paste your own repository in the box below and FoxPlug writes the same weekly update for your product. This page is an automatic summary of the releases, pull requests and commits in github.com/kedacore/keda; Keda did not write it and does not use or endorse FoxPlug. Every line links to the change it describes.

Get a weekly update like this for your product, free

Or Use it as a GitHub Action

See the update FoxPlug writes for keda, free

Week of November 2, 2026

What shipped

Why it matters

This week brought fixes that change behaviour for people running scalers in production: Elasticsearch and Elastic Forecast can now use API keys on self-hosted clusters, IPv6 endpoints work in the Metrics API scaler, and ArangoDB finally honours its connection limit. Several correctness fixes also landed in metrics collection, webhook readiness, and Splunk Observability output handling, which reduces flakiness and resource leakage under sustained polling.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: OpenTelemetry metric data race fixed, webhook readiness now waits for the server, API key auth for Elasticsearch and Elastic Forecast on self-hosted clusters, IPv6 addresses bracketed in Metrics API scaler URLs, and connectionLimit honoured by the ArangoDB scaler.

Post for LinkedIn

This week in KEDA: a data race in OpenTelemetry metric callbacks was fixed, and the admission webhook readiness check now waits for the webhook server to start. The Elasticsearch and Elastic Forecast scalers accept API keys when connecting through addresses, so self-hosted clusters can authenticate without cloudID. The Metrics API scaler brackets IPv6 endpoint addresses in aggregated URLs, and the ArangoDB scaler now applies connectionLimit as the maximum connections per host. The Splunk Observability scaler drains all SignalFlow computation outputs, and scaler unit tests got cross-platform and isolation fixes.

Week of October 19, 2026

What shipped

Why it matters

This week brings fixes that change how scaling behaves when things are not straightforward: unchanged streamed specs are no longer reprocessed, InfluxDB query failures are reported rather than treated as an empty result, and very large metric values no longer wrap to a negative number before rounding up. The new Metrics API scaler option gives operators control over how missing ready endpoints are treated. Keeping the devcontainer and toolchain Go versions aligned helps development match CI.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: unchanged streamed metric specs are skipped, InfluxDB scaler errors are surfaced instead of read as empty, ceilToInt64 clamps to avoid overflow, and the Metrics API scaler adds a zeroOnNoReadyEndpoints option.

Post for LinkedIn

This week in the project: streamed metric specs that match the previous cached update are now skipped, avoiding repeated HPA and status processing. The InfluxDB scaler surfaces query errors instead of reporting them as an empty result, so scaling no longer proceeds on a zero metric. ceilToInt64 now clamps to math.MaxInt64 instead of wrapping on very large float inputs. The Metrics API scaler adds an opt-in zeroOnNoReadyEndpoints option, default false, for aggregateFromKubeServiceEndpoints. Build instructions also keep the devcontainer and keda-toolchain images on the same Go version.

Week of October 12, 2026

What shipped

Why it matters

This week fixes a real correctness gap in the GitHub Runner scaler, where runs with more than 100 jobs could hide still-queued work, and makes the NSQ scaler usable against nsq versions before 1.0. It also removes a deprecated InfluxDB trigger field, so anyone still setting it should drop it. The remaining changes cut fixed sleeps from the test suite, which should reduce flaky CI runs.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: GitHub Runner scaler follows job pagination past the first 100 jobs, NSQ scaler speaks the V1 HTTP API for nsq below 1.0, InfluxDB drops the deprecated authToken metadata, and CI tests poll instead of sleeping.

Post for LinkedIn

This week in KEDA: the GitHub Runner scaler now follows pagination for workflow run jobs, so runs with more than 100 jobs no longer hide queued or in-progress work beyond the first page. The NSQ scaler requests the V1 HTTP API format, supporting nsq versions below 1.0. The deprecated authToken trigger metadata is removed from the InfluxDB scaler. On the test side, several in-cluster scaler tests and helper loops now poll their dependencies instead of sleeping between checks, which tackles flakiness in OpenShift CI.

Week of September 21, 2026

What shipped

Why it matters

KEDA 2.21.0 introduces important security and stability improvements with breaking changes requiring review before upgrade. Multiple resource leak fixes and data race corrections prevent production issues with connection pooling, certificate rotation, and context lifecycle management. Enhanced metrics support and kafka scaler accuracy improvements increase visibility and correct scaling behavior for common use cases.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

KEDA 2.21.0 is out. This release includes breaking changes and fixes for gRPC connection leaks, certificate rotation data races, and Kafka scaler accuracy. Review the upgrade guide before updating.

Post for LinkedIn

KEDA 2.21.0 is available. This release addresses critical stability issues including data race fixes in mTLS certificate rotation, proper gRPC connection cleanup, and corrected Kafka lag calculation for retention-deleted messages. Three breaking changes are documented in the upgrade guide. Enhanced gRPC client metrics provide better visibility into scaler performance.

Week of September 14, 2026

What shipped

Why it matters

This week adds significant authentication improvements across multiple scalers with OAuth2 and Kerberos support, introduces a new Azure Cosmos DB scaler, and resolves critical lifecycle and timeout issues that affected reliability. The fixes to scaling behavior and configuration validation improve operator stability.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

KEDA this week: Azure Cosmos DB scaler, OAuth2 & Kerberos auth support, Kubernetes API timeouts, and fixes for Azure Pipelines and Vault integration.

Post for LinkedIn

KEDA updates this week bring new capabilities and fixes across multiple areas. The new Azure Cosmos DB Change Feed Scaler translates C# functionality to native Go. Prometheus, Kafka, and Metrics API scalers gain enhanced authentication support with OAuth2 and Kerberos. Critical reliability fixes address Kubernetes API timeouts, Azure Pipelines scaling, and token renewal issues.

Weeks with too little public activity are left out rather than filled in. Last updated 2026-11-10.

Is this your repo? Ask us to remove this page.