What we're building next. Completed items are included to show the journey — the platform has come a long way.
Share an Ingress with other services by configuring a sub-path for Kwirth.
One-command installation with full Ingress and TLS support.
Issue, revoke, and expire API keys from the UI.
Restart Kwirth from the UI — with latest image tag this auto-updates.
Extensibility subsystem for hot-reloadable front-end channels and data providers.
Sign in with Google, GitLab (cloud & self-managed) and GitHub (cloud & Enterprise) via OIDC / OAuth2 — installable connectors, hardened with PKCE + back-channel.
Replace proprietary API keys with JWT scopes matching the Kwirth VIEW system.
Add GatewayClass and HTTPRoute resources to the Helm chart.
Prepare the project for donation to the Cloud Native Computing Foundation.
Configure channels and providers through environment variables (KWIRTH_CHANNEL_<id>_<var>).
Choose to receive logs since pod start, previous logs, or only from now on.
Process log messages for alerting without storing them.
Mask lines containing sensitive keywords (password, token, email) automatically.
Show a clear in-viewer indicator when the WebSocket connection drops or errors occur.
Open, resize, pin, and minimize multiple artifact detail windows simultaneously.
Open a terminal session directly to a cluster node via KwirthWork.
Filter the resource tree by namespace, node, and other criteria.
Lightweight native desktop build powered by Tauri — smaller binary and lower memory footprint than the Electron build. Available for Windows and Linux.
For large clusters, limit Magnify sync to selected namespaces to reduce memory and API pressure.
Let users assign Kinds to custom tree sections (e.g. show HTTPRoute under Network instead of Custom).
After editing a container's filesystem via File Manager, push the modified image back to a registry using nerdctl.
Continuous cluster analysis with anomaly detection, explanations, and recommendations.
Test and tune prompts and tool configurations from the UI without restarting.
Version trigger and LLM configurations — enable/disable specific versions independently.
Send Pinocchio findings out through Kwirth senders (email, Teams, and more) — not just the in-app log.
Let Pinocchio apply suggested changes via Kubernetes mutating webhook or IaC (ArgoCD, Flux, Helm).
Expose clear traces of what tools the LLM invoked and why, for easier prompt debugging.
Advanced scan configuration — inclusions, exclusions, per-channel settings.
Ingest traces, metrics, and logs via standard OTLP/HTTP into Kwirth channels.
Stream Kafka events into Kwirth for unified infrastructure and application observability.
Embed Kwirth log streams and metrics directly in Grafana dashboards.
CPU, memory, and status via WebSocket streaming.
Trigger alerts when CPU or memory cross configurable thresholds — not just log patterns.
Aggregate data from multiple clusters and sources into a single unified channel.