ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Developer tools · #kubeflow
Kubeflow Pipelines v2 reuses task outputs when a task's component signature and inputs match a previous successful execution. This works well until a component's container image is rebuilt under the same tag: the cache key does not appear to change, so stale outputs can be silently reused even though the underlying code changed. The deployment in question us
MLMD schema mismatch after Kubeflow Pipelines database restore Restoring the ML Metadata (MLMD) MySQL database for Kubeflow Pipelines should recover pipeline execution history. The restored schema must match the version expected by the deployed KFP control plane, or the UI may fail to reconcile executions. A restored MLMD store can also reference artifacts i
The Kubeflow Pipelines UI offers basic arrow‑key traversal of the pipeline graph, but focus management for nested components remains incomplete. Custom HTML elements represent nodes without ARIA labels, preventing screen readers from identifying node types or connections. The current release (1.8.x) includes a focus trap in modal dialogs, yet the editor itse
503 Service Unavailable in Kubeflow after upgrade: Istio routing misconfiguration During a recent upgrade of a Kubeflow deployment that uses Kustomize manifests, the cluster started returning 503 Service Unavailable responses from the Ingress gateway. The issue appears to be linked to Istio’s virtual service and destination rule definitions that were regener
We run a standard Kubeflow Manifests deployment where pipeline run metadata lives in the MySQL metadata database and run artifacts live in an external object storage bucket. We are designing a restore procedure that covers both, since a control-plane backup alone does not protect the artifacts. The unresolved decision is verification. After restoring the met
The goal is to determine whether Velero‑based filesystem backups or CSI‑driven volume snapshots provide a reliable, application‑consistent restore path for Kubeflow Pipelines artifact stores (MLflow and MinIO) without triggering the 'timeout waiting for pod to be ready' error during workflow‑controller pod recovery. The uncertainty stems from Velero restore
Goal To evaluate how the default single‑threaded inference server (concurrency: 1) in Kubeflow KServe affects request latency when multiple inference requests are sent simultaneously. Constraints The inference server runs without the predictor.serving.kserve.io/concurrency annotation, inheriting the built‑in limit. The Kubernetes HPA does not scale pods auto
This question addresses the optimal token lifetime configuration for Kubeflow's least-privilege authentication model, where security requirements conflict with operational overhead. Kubeflow's native RBAC provides granular resource-level permissions but does not natively enforce token expiration, relying instead on external identity provider token lifetimes.
Kubeflow Pipelines and KServe: Zero-Downtime Model Migration Boundary Kubeflow Pipelines compiles workflow definitions into Argo Workflow YAML and executes PipelineRuns as immutable instances. KServe manages InferenceService deployments and relies on Istio for canary rollouts and traffic splitting. The integration boundary is the handoff from a completed tra
kubernetes