Skip to content
Ironflow Docs

Ironflow v0.17.0

Ironflow v0.17.0 is a major dashboard and deployment release. The workspace picker is now unified — project and environment selection merged into a single dropdown. The circuit breaker gains cluster-wide persistence via NATS KV. And the operator story gets a significant upgrade: multi-tenant Helm deployments, Hetzner load balancer support with Traefik ingress, and a fully unified SRE monitoring stack built into the Helm chart.

Features

Section titled “Features”

Dashboard

Section titled “Dashboard”
  • Unified workspace picker — project and environment merged into one dropdown, workspace context preserved across navigation (#369)
  • Environment colors — environments support a selectable color for visual distinction (#371)
  • Org picker and platform sentinel filter — new org switcher in the header; platform sentinel filtering on Audit Log (#378)
  • Multi-select event type filter on Audit Log (#379)
  • Expandable function rows and run detail enrichments — richer inline detail on the Functions and Runs pages (#364)
  • Copy-to-clipboard on error panels — one-click copy of run and step error messages (#368)
  • Brand guide enforcement — semantic color tokens and consistent typography across all dashboard pages (#367)

Platform

Section titled “Platform”
  • Circuit breaker state in NATS KV — breaker state persists across server restarts and is shared cluster-wide via NATS KV watch (#359)
  • ~/.ironflow/ fallback for license keys — the binary now checks ~/.ironflow/license.key and ~/.ironflow/license_pub.pem as a fallback (#385)
  • SRE runtime hardening — NATS resilience improvements, dead-letter queue, and expanded observability (#333)

Deployment & Operations

Section titled “Deployment & Operations”
  • Multi-tenant Helm deployment — namespace-per-tenant isolation with network policies and resource quotas (#342)
  • Enterprise Edition deployment support — EE-specific Helm configuration and provisioning fixes (#340, #377)
  • k3d local provider — ironflow provision now supports k3d for local cluster development (#338)
  • Hetzner load balancer with Traefik ingress — production-grade Hetzner deployments with TLS termination (#344)
  • Unified Grafana dashboards and alerts in Helm chart — monitoring bundled into the chart; NATS monitoring, kube-state-metrics, and SRE dashboards added (#352, #346)
  • kube-prometheus-stack on all deployment templates — monitoring enabled by default across all four deployment sizes (#336)
  • make helm-redeploy target — faster local k3d iteration (#354)

AI Skills

Section titled “AI Skills”
  • ironflow-platform skill — new skill for cluster ops, deployment, and provisioning from AI assistants (#348)
  • Interactive setup skill + CLI-first pattern — ironflow-setup now interactive; all skills updated to CLI-first with full 27/27 MCP tool coverage (#347, #357)

Fixes

Section titled “Fixes”
  • Resolve dashboard 401 on project/environment switch (#381)
  • Refresh dashboard data on project/env switch; show full event metadata (#370)
  • Add missing Vite dev proxy rules for ConnectRPC and WebSocket (#380)
  • Fix workspace picker menu opening to the left of the viewport (#372)
  • Resolve test flakes in E2E cancel, race detector, and CI NATS port conflicts (#384)
  • Fix CNPG recovery bootstrap and per-tenant backup gaps in Helm (#350)
  • Raise all packages above 80% test coverage gate (#349)
  • Fix Helm test SIGPIPE failures and expand template test coverage (#355)
  • Improve Hetzner teardown cleanup (#345)