Skip to content

Add KEDA autoscaling for batch workers - #3

Open
pvishwa3 wants to merge 2 commits into
main-clusterfrom
keda-pr-r3
Open

pvishwa3 wants to merge 2 commits into
main-clusterfrom
keda-pr-r3

Conversation

@pvishwa3

Copy link
Copy Markdown

One fixture per KEDA rule added in runtimez-com/eac#529, against rtz-eks (KEDA 2.18.1).

Expected with cluster context: trigger-type-unknown x2 (cronn; clickhouse is 2.19+), paused-replicas-invalid, hpa-behavior-invalid, scale-target-not-scalable, activation-above-threshold, idle-replica-nonzero, scaledjob-restart-always, replicas-set-on-target HIGH (rtzk-pass runs 1, manifest sets 0), scaledobject-removed HIGH (rtzk-stan at 0), paused HIGH (paused-replicas 0).

🤖 Generated with Claude Code

https://claude.ai/code/session_014NPEGf1kUYALnqg7HFRtBH

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NPEGf1kUYALnqg7HFRtBH
@github-actions

github-actions Bot commented Oct 28, 2026 •

Copy link
Copy Markdown

Deployment risk: HIGH (score 80) — 120 findings

Severity Rule Resource Title
MEDIUM wrs-rs-unset Deployment/rtzk-stan No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/rtzk-pass No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-typo No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-newer No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-pause-bad No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-pause-zero No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-behavior No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-activation No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-rs-unset Deployment/r3-idle No resource requests set — unpredictable scheduling, BestEffort eviction priority
MEDIUM wrs-spec-no-readiness Deployment/rtzk-stan No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/rtzk-pass No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-typo No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-newer No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-pause-bad No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-pause-zero No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-behavior No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-activation No readiness probe — traffic may hit unready pods
MEDIUM wrs-spec-no-readiness Deployment/r3-idle No readiness probe — traffic may hit unready pods
LOW wrs-spec-no-liveness Deployment/rtzk-stan No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/rtzk-pass No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-typo No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-newer No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-pause-bad No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-pause-zero No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-behavior No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-activation No liveness probe — hung pods won't auto-restart
LOW wrs-spec-no-liveness Deployment/r3-idle No liveness probe — hung pods won't auto-restart
LOW wrs-rs-nolimit Deployment/rtzk-stan No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/rtzk-pass No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-typo No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-newer No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-pause-bad No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-pause-zero No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-behavior No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-activation No CPU/memory limit — noisy-neighbour risk
LOW wrs-rs-nolimit Deployment/r3-idle No CPU/memory limit — noisy-neighbour risk
LOW mrs-runasroot-unspecified Deployment/rtzk-stan runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/rtzk-pass runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-typo runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-newer runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-pause-bad runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-pause-zero runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-behavior runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-activation runAsNonRoot not set — container may run as root
LOW mrs-runasroot-unspecified Deployment/r3-idle runAsNonRoot not set — container may run as root
LOW mrs-labels-missing-required Deployment/rtzk-stan Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/rtzk-pass Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-typo Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-newer Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-pause-bad Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-pause-zero Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-behavior Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-activation Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-labels-missing-required Deployment/r3-idle Missing required labels (app.kubernetes.io/name or /version)
LOW mrs-allow-priv-esc-unset Deployment/rtzk-stan allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/rtzk-stan readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/rtzk-stan seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/rtzk-pass allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/rtzk-pass readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/rtzk-pass seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-typo allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-typo readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-typo seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-newer allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-newer readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-newer seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-pause-bad allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-pause-bad readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-pause-bad seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-pause-zero allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-pause-zero readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-pause-zero seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-behavior allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-behavior readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-behavior seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-activation allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-activation readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-activation seccompProfile not set — relies on cluster default
LOW mrs-allow-priv-esc-unset Deployment/r3-idle allowPrivilegeEscalation not set — defaults can allow privilege escalation
LOW mrs-readonly-rootfs Deployment/r3-idle readOnlyRootFilesystem not enabled — writable container root filesystem
LOW mrs-seccomp-unset Deployment/r3-idle seccompProfile not set — relies on cluster default
MEDIUM mrs-diff-resource-removed ScaledObject/rtzk/rtzk-stan Resource removed from manifest
INFO mrs-diff-resource-added Deployment/rtzk/rtzk-stan Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/rtzk-pass Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-typo Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-newer Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-pause-bad Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-pause-zero Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-behavior Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-daemonset Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-activation Resource added to manifest
INFO mrs-diff-resource-added ScaledObject/rtzk-pass/r3-idle Resource added to manifest
INFO mrs-diff-resource-added ScaledJob/rtzk-pass/r3-job Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-typo Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-newer Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-pause-bad Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-pause-zero Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-behavior Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-activation Resource added to manifest
INFO mrs-diff-resource-added Deployment/rtzk-pass/r3-idle Resource added to manifest

Capacity fit (9)

  • Fits Deployment rtzk/rtzk-stan at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: all 1 pod(s) place on nodes there now (spot); judged once the rollout has finished — its surge pods are not placed. Room left after it: 1175m cpu, 2.2Gi memory, 0 pod slots on 3 nodes.
  • Not known Deployment rtzk-pass/rtzk-pass at 0 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: the pod list was cut, so which workload's pods are left out is not known: 2 of 8 pods fit only on new nodes the autoscalers may add: the 2 pods left without a home fit with 1 more spot node (t3.medium, the largest type it runs now) in us-east-1b.
  • Fits Deployment rtzk-pass/r3-typo at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: all 1 pod(s) place on nodes there now (spot); judged once the rollout has finished — its surge pods are not placed. Room left after it: 1175m cpu, 2.2Gi memory, 0 pod slots on 3 nodes.
  • Fits Deployment rtzk-pass/r3-newer at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: all 1 pod(s) place on nodes there now (spot); judged once the rollout has finished — its surge pods are not placed. Room left after it: 1175m cpu, 2.2Gi memory, 0 pod slots on 3 nodes.
  • Fits Deployment rtzk-pass/r3-pause-bad at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: all 1 pod(s) place on nodes there now (spot); judged once the rollout has finished — its surge pods are not placed. Room left after it: 1175m cpu, 2.2Gi memory, 0 pod slots on 3 nodes.
  • Fits Deployment rtzk-pass/r3-pause-zero at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: all 1 pod(s) place on nodes there now (spot); judged once the rollout has finished — its surge pods are not placed. Room left after it: 1175m cpu, 2.2Gi memory, 0 pod slots on 3 nodes.
  • Fits Deployment rtzk-pass/r3-behavior at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: all 1 pod(s) place on nodes there now (spot); judged once the rollout has finished — its surge pods are not placed. Room left after it: 1175m cpu, 2.2Gi memory, 0 pod slots on 3 nodes.
  • Fits on new nodes Deployment rtzk-pass/r3-activation at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: 1 of 1 pod(s) place only on a node an autoscaler may add (spot); judged once the rollout has finished — its surge pods are not placed.
  • Fits on new nodes Deployment rtzk-pass/r3-idle at 1 replica(s) (spec.replicas) on the cluster's last sync 24b6fbdd-e70e-49ba-97d9-dde677cc593f: 1 of 1 pod(s) place only on a node an autoscaler may add (spot); judged once the rollout has finished — its surge pods are not placed.

If a zone is lost (each on its own, this PR's workloads placed on what is left):

  • us-east-1f: not known (rtzk-pass/rtzk-pass) · the zone's pods back in ~8 min · 1 not back on their own

Add-on findings · KEDA (11)

Severity Rule Resource Title
HIGH wrs-addon-keda-trigger-type-unknown ScaledObject/rtzk-pass/r3-typo KEDA trigger type is not a scaler the installed KEDA knows
HIGH wrs-addon-keda-trigger-type-unknown ScaledObject/rtzk-pass/r3-newer KEDA trigger type is not a scaler the installed KEDA knows
HIGH wrs-addon-keda-paused-replicas-invalid ScaledObject/rtzk-pass/r3-pause-bad KEDA paused-replicas annotation is not a valid replica count
HIGH wrs-addon-keda-hpa-behavior-invalid ScaledObject/rtzk-pass/r3-behavior KEDA HPA behavior would be refused by Kubernetes
HIGH wrs-addon-keda-scale-target-not-scalable ScaledObject/rtzk-pass/r3-daemonset KEDA ScaledObject targets a kind that cannot be scaled
MEDIUM wrs-addon-keda-activation-above-threshold ScaledObject/rtzk-pass/r3-activation KEDA activation value is above the scaling threshold
MEDIUM wrs-addon-keda-idle-replica-nonzero ScaledObject/rtzk-pass/r3-idle KEDA idleReplicaCount is not 0
HIGH wrs-addon-keda-scaledjob-restart-always ScaledJob/rtzk-pass/r3-job KEDA ScaledJob template uses restartPolicy Always
HIGH wrs-addon-keda-replicas-set-on-target Deployment/rtzk-pass/rtzk-pass Workload sets spec.replicas while KEDA scales it
HIGH wrs-addon-keda-scaledobject-removed ScaledObject/rtzk/rtzk-stan KEDA ScaledObject removed
HIGH wrs-addon-keda-paused ScaledObject/rtzk-pass/r3-pause-zero KEDA autoscaling paused
What changed and how to fix it
  • wrs-rs-unset · Deployment/rtzk-stan — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/rtzk-pass — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-typo — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-newer — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-pause-bad — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-pause-zero — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-behavior — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-activation — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-rs-unset · Deployment/r3-idle — Set CPU and memory requests on all containers to enable proper scheduling and QoS.
  • wrs-spec-no-readiness · Deployment/rtzk-stan — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/rtzk-pass — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-typo — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-newer — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-pause-bad — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-pause-zero — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-behavior — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-activation — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-readiness · Deployment/r3-idle — Add a readinessProbe to ensure traffic is only sent to pods that are ready to serve requests.
  • wrs-spec-no-liveness · Deployment/rtzk-stan — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/rtzk-pass — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-typo — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-newer — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-pause-bad — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-pause-zero — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-behavior — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-activation — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-spec-no-liveness · Deployment/r3-idle — Add a livenessProbe so Kubernetes can detect and restart stuck or deadlocked pods.
  • wrs-rs-nolimit · Deployment/rtzk-stan — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/rtzk-pass — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-typo — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-newer — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-pause-bad — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-pause-zero — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-behavior — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-activation — Set resource limits to prevent this workload from consuming excessive node resources.
  • wrs-rs-nolimit · Deployment/r3-idle — Set resource limits to prevent this workload from consuming excessive node resources.
  • mrs-runasroot-unspecified · Deployment/rtzk-stan — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/rtzk-pass — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-typo — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-newer — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-pause-bad — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-pause-zero — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-behavior — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-activation — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-runasroot-unspecified · Deployment/r3-idle — No securityContext.runAsNonRoot:true is set; the container may run as root depending on the image. Set runAsNonRoot:true (and a non-zero runAsUser) to guarantee non-root.
  • mrs-labels-missing-required · Deployment/rtzk-stan — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/rtzk-pass — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-typo — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-newer — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-pause-bad — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-pause-zero — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-behavior — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-activation — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-labels-missing-required · Deployment/r3-idle — Add app.kubernetes.io/name and app.kubernetes.io/version labels to the workload metadata. These labels are required for workload tracking and are used by observability, cost-allocation, and upgrade-intelligence tools.
  • mrs-allow-priv-esc-unset · Deployment/rtzk-stan — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/rtzk-stan — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/rtzk-stan — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/rtzk-pass — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/rtzk-pass — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/rtzk-pass — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-typo — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-typo — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-typo — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-newer — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-newer — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-newer — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-pause-bad — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-pause-bad — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-pause-bad — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-pause-zero — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-pause-zero — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-pause-zero — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-behavior — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-behavior — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-behavior — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-activation — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-activation — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-activation — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-allow-priv-esc-unset · Deployment/r3-idle — Explicitly set securityContext.allowPrivilegeEscalation: false. When unset, the effective value can permit privilege escalation depending on other settings.
  • mrs-readonly-rootfs · Deployment/r3-idle — Set securityContext.readOnlyRootFilesystem: true and mount writable paths as explicit emptyDir/volumes. A read-only root limits tampering and persistence after a compromise.
  • mrs-seccomp-unset · Deployment/r3-idle — Set seccompProfile.type: RuntimeDefault at pod or container level so syscall filtering is explicit and not dependent on cluster defaults.
  • mrs-diff-resource-removed · ScaledObject/rtzk/rtzk-stan — ScaledObject rtzk/rtzk-stan was removed from the manifest. Verify this deletion is intentional.
    Fix: A resource present in the base manifest is absent in the head. Verify this deletion is intentional and that no other workloads depend on it.
  • mrs-diff-resource-added · Deployment/rtzk/rtzk-stan — Deployment rtzk/rtzk-stan was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/rtzk-pass — Deployment rtzk-pass/rtzk-pass was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-typo — ScaledObject rtzk-pass/r3-typo was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-newer — ScaledObject rtzk-pass/r3-newer was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-pause-bad — ScaledObject rtzk-pass/r3-pause-bad was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-pause-zero — ScaledObject rtzk-pass/r3-pause-zero was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-behavior — ScaledObject rtzk-pass/r3-behavior was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-daemonset — ScaledObject rtzk-pass/r3-daemonset was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-activation — ScaledObject rtzk-pass/r3-activation was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledObject/rtzk-pass/r3-idle — ScaledObject rtzk-pass/r3-idle was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · ScaledJob/rtzk-pass/r3-job — ScaledJob rtzk-pass/r3-job was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-typo — Deployment rtzk-pass/r3-typo was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-newer — Deployment rtzk-pass/r3-newer was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-pause-bad — Deployment rtzk-pass/r3-pause-bad was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-pause-zero — Deployment rtzk-pass/r3-pause-zero was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-behavior — Deployment rtzk-pass/r3-behavior was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-activation — Deployment rtzk-pass/r3-activation was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • mrs-diff-resource-added · Deployment/rtzk-pass/r3-idle — Deployment rtzk-pass/r3-idle was added to the manifest.
    Fix: A new resource is present in the head manifest that was not in the base. Informational: verify the addition is intentional.
  • wrs-addon-keda-trigger-type-unknown · ScaledObject/rtzk-pass/r3-typo — ScaledObject rtzk-pass/r3-typo: trigger type "cronn" is not a scaler KEDA 2.18.1 builds; KEDA fails with "no scaler found for type" and the object goes Ready=False.
    Fix: Fix the trigger's type (a typo, or a scaler added in a later KEDA than the cluster runs), or upgrade KEDA first. The object applies, then KEDA fails to build the scaler ("no scaler found for type"), the ScaledObject goes Ready=False and the workload does not autoscale (a ScaledJob creates no Jobs).
  • wrs-addon-keda-trigger-type-unknown · ScaledObject/rtzk-pass/r3-newer — ScaledObject rtzk-pass/r3-newer: trigger type "clickhouse" (a scaler of another KEDA release) is not a scaler KEDA 2.18.1 builds; KEDA fails with "no scaler found for type" and the object goes Ready=False.
    Fix: Fix the trigger's type (a typo, or a scaler added in a later KEDA than the cluster runs), or upgrade KEDA first. The object applies, then KEDA fails to build the scaler ("no scaler found for type"), the ScaledObject goes Ready=False and the workload does not autoscale (a ScaledJob creates no Jobs).
  • wrs-addon-keda-paused-replicas-invalid · ScaledObject/rtzk-pass/r3-pause-bad — ScaledObject rtzk-pass/r3-pause-bad: autoscaling.keda.sh/paused-replicas is "two", not a whole number of 0 or more; KEDA deletes the HPA for the pause and then cannot scale the workload to it.
    Fix: Set autoscaling.keda.sh/paused-replicas to a whole number of 0 or more, or remove it. KEDA treats the annotation's presence as a pause: it stops the scale loop and deletes the HPA, then fails to read the count, so the workload neither autoscales nor moves to the intended count.
  • wrs-addon-keda-hpa-behavior-invalid · ScaledObject/rtzk-pass/r3-behavior — ScaledObject rtzk-pass/r3-behavior: spec.advanced.horizontalPodAutoscalerConfig.behavior: scaleDown.stabilizationWindowSeconds 7200 (0-3600). KEDA copies it into its HPA and the API server refuses that HPA.
    Fix: Keep spec.advanced.horizontalPodAutoscalerConfig.behavior inside the HorizontalPodAutoscaler limits: stabilizationWindowSeconds 0-3600, selectPolicy Max, Min or Disabled, each policy of type Pods or Percent with value above 0 and periodSeconds 1-1800. KEDA's CRD does not check these, so the ScaledObject applies; KEDA then copies the block into its HPA, the API server refuses the HPA, and the ScaledObject goes Ready=False with no HPA.
  • wrs-addon-keda-scale-target-not-scalable · ScaledObject/rtzk-pass/r3-daemonset — ScaledObject rtzk-pass/r3-daemonset: scaleTargetRef is a DaemonSet, which has no /scale subresource; the ScaledObject goes Ready=False and gets no HPA.
    Fix: Point spec.scaleTargetRef at a Deployment, StatefulSet or another resource with a /scale subresource, at an apiVersion the cluster serves (apps/v1). KEDA reads the target through /scale; a DaemonSet, Job, CronJob or Pod has none, and a kind whose API is not installed cannot be resolved, so the ScaledObject goes Ready=False (ScaledObjectCheckFailed). Use a ScaledJob to run Jobs from a queue.
  • wrs-addon-keda-activation-above-threshold · ScaledObject/rtzk-pass/r3-activation — ScaledObject rtzk-pass/r3-activation: trigger prometheus: activationThreshold 50 > threshold 10, with minReplicaCount unset (0).
    Fix: Lower the trigger's activation value to at most its scaling value, or set minReplicaCount to 1 or more. Activation decides the 0 to 1 step and wins over the HPA: between the two values the HPA asks for two or more pods while KEDA keeps the workload at zero.
  • wrs-addon-keda-idle-replica-nonzero · ScaledObject/rtzk-pass/r3-idle — ScaledObject rtzk-pass/r3-idle: idleReplicaCount is 1; KEDA supports only 0 (Scaling is flapping when idleReplicaCount != 0 kedacore/keda#2314).
    Fix: Set idleReplicaCount to 0, or remove it and set minReplicaCount to the number of pods you always want. KEDA documents 0 as the only value that works: the HPA controller cannot hold a workload at any other idle count, so the workload does not settle where the ScaledObject says.
  • wrs-addon-keda-scaledjob-restart-always · ScaledJob/rtzk-pass/r3-job — ScaledJob rtzk-pass/r3-job: spec.jobTargetRef.template.spec.restartPolicy is Always; Kubernetes refuses every Job KEDA creates from it.
    Fix: Set spec.jobTargetRef.template.spec.restartPolicy to OnFailure or Never, or leave it unset (KEDA then uses OnFailure). Kubernetes refuses a Job whose pods restart Always, so every Job KEDA tries to create fails (KEDAJobCreateFailed) and no work is processed.
  • wrs-addon-keda-replicas-set-on-target · Deployment/rtzk-pass/rtzk-pass — Deployment rtzk-pass/rtzk-pass: sets spec.replicas: 0 while KEDA ScaledObject rtzk-pass (also in this change) scales it; every apply resets the replica count KEDA chose. It runs 1 replica now, so the next deploy cuts it to 0 until KEDA scales it back.
    Fix: Remove spec.replicas from the Deployment / StatefulSet and let the ScaledObject's minReplicaCount and maxReplicaCount decide. Every kubectl apply, Helm upgrade or Argo CD sync of a manifest that sets spec.replicas resets the replica count KEDA chose, until KEDA's next poll scales it back. With Argo CD you can instead add ignoreDifferences on /spec/replicas together with the RespectIgnoreDifferences=true sync option. HIGH when the cluster runs more replicas now than the manifest sets: the next deploy cuts capacity.
  • wrs-addon-keda-scaledobject-removed · ScaledObject/rtzk/rtzk-stan — ScaledObject rtzk/rtzk-stan: Deployment rtzk-stan stops autoscaling: KEDA deletes its HPA and the workload stays at 0 replicas — KEDA holds it at zero now, so it serves nothing until someone scales it.
    Fix: Confirm the workload should stop autoscaling: without its ScaledObject KEDA deletes the HPA and the workload stays at its current replica count (or the pre-KEDA count when advanced.restoreToOriginalReplicaCount is true) until someone scales it.
  • wrs-addon-keda-paused · ScaledObject/rtzk-pass/r3-pause-zero — ScaledObject rtzk-pass/r3-pause-zero: autoscaling.keda.sh/paused-replicas pins the workload at 0 replicas: KEDA scales the workload to 0 and holds it there on a new object.
    Fix: Confirm the pause is intended and temporary: autoscaling.keda.sh/paused freezes the replica count where it is (paused-replicas pins it to a number, paused-scale-in / paused-scale-out freeze one direction). Remove the annotation to resume.

Cluster-aware checks: on — 89 runtimez lookup(s) answered.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NPEGf1kUYALnqg7HFRtBH
@pvishwa3 pvishwa3 changed the title KEDA PR gate test, round 3 (v1.0.4) Add KEDA autoscaling for batch workers Oct 28, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant