Skip to content
Back to blog

What's coming in v0.6.0

v0.6.0 ships at the end of June, and it has a theme: fewer places to look. Notifications come to you instead of waiting in an inbox. Four Operations tabs collapse into one view. And the catalog stops pretending every Kubernetes workload is a “service”.

Alerts that find you

Notifications can go where your team already lives: Slack, email, or any webhook. Paste a Slack or webhook secret and Shoehorn stores it encrypted, or point the subscription at a secret you’ve already configured. Team admins get a Notifications tab with every subscription in one table, a test-send button to check the wiring before you trust it, and one-click resume for anything paused.

One Operations page

Finding a struggling workload used to mean hopping between Overview, Workloads, and GitOps. Those tabs are gone. In their place: a single Resources view with every workload across all your clusters, and a detail drawer on each one.

Filters live in the URL. Filter down to “failing deployments in production, cluster east”, copy the address bar, drop it in Slack. Your teammate opens exactly what you’re looking at.

There’s also a new Signals column showing what’s firing on each resource. A healthy row stays empty.

And one genuinely new trick: Helm release detection. The Kubernetes agent can now spot the Helm releases running in your clusters and show which workloads belong to which release. Ever stared at twelve deployments wondering which chart put them there? That’s the question this answers. It’s off by default; enable it in the agent’s Helm chart or through the Terraform module.

A dashboard that shows direction

A number on a KPI card tells you where you are. It doesn’t tell you if you’re sliding. The home dashboard cards now carry a 14-day sparkline and trend, built from daily snapshots that start the day you upgrade.

New on the dashboard: a Knowledge Risk tile. It counts the repositories that depend on a single maintainer, the ones that turn into a problem the day that person goes on holiday. Click through for the full list.

The catalog gets honest

Until now, every Kubernetes-imported entity showed up as a generic “service”. A StatefulSet, a CronJob, a Deployment: all the same gray label. In v0.6.0 they show their real workload kind, and if you want your own classification, set the shoehorn.dev/type annotation and Shoehorn uses that instead.

Other catalog changes:

  • A Tier column shows each service’s criticality (Critical, High, Medium, Low) as a compact set of bars, and you can sort by it. Manifests still using Bronze, Silver, Gold, or Platinum keep working; they map onto the new names.

  • Docs can attach directly to a team. Add related_teams to a doc’s frontmatter and it lands on that team’s Documentation tab. Team charters, on-call policies, and ways of working finally have a home that isn’t a single service.

  • Pressing back from a service’s detail page keeps your filters, search, and scroll position. A small fix for a daily annoyance.

Governance that knows what’s urgent

The Governance page now sorts by urgency, with overdue work at the top where you can’t ignore it. Resolve or dismiss several actions at once, filter to overdue, critical, or high priority, and find resolved items in their own group instead of mixed into the pile.

Visual refresh

v0.6.0 marks the beginning of our new visual refresh. First up is: Operations, Entity pages, the Dashboard, Insights, Code, and Security. A cleaner layout and more metadata.

Stop counting nodes

Free no longer caps node count. Licensing now counts active Kubernetes agents, and Free covers one active agent (one cluster) with as many nodes as you want under it. Your existing license converts automatically. There’s nothing to do.

Installs that fail loudly or not at all

If you’ve ever had a Helm install go through and then quietly misbehave, this release is for you. The chart now fails fast with a clear error when required values are missing, instead of deploying something broken. Installs on clusters without a default StorageClass work. The RBAC sync job renders correctly. Valkey actually enforces the password you gave it.

When a release moves to a newer version of the bundled search engine, you can migrate the existing index in place instead of rebuilding it from scratch.

Security

We updated our hardened base images and patched Go and web dependencies. The agent push API also got stricter: dashboard URLs that aren’t plain http or https are rejected before they can ever reach the UI.

Going away

  • The Pipelines page under Code.
  • The standalone GitOps fleet page. The ArgoCD and Flux sync view lives on each entity’s GitOps tab for now.
  • The separate netobserver component. Its job moved into the regular Kubernetes agent.

Changelog

The full changelog will be published at release here.

In other news