LoginSign Up
← Back to all sparks
W

Windmill

INFRA · APIS
Velocity6.3

Open-source developer platform to build internal tools and workflows from scripts

www.windmill.dev ↗

Windmill hardens its runtime: daemonless containers, SSH execution, dev/prod workspaces.

workflow-enginecontainer-runtimemulti-tenant-isolationenterprise-opsobservability
◆Current state
Windmill is an open-source developer platform that runs scripts, flows, and apps as a workflow and internal-tools engine. Its changelog is a genuine, dense stream of runtime and operations features. The current cluster centers on execution flexibility (containers, remote SSH, SQL S3 inputs), environment safety (dev/prod workspaces), and enterprise observability and governance (OTEL tracing, audit-log export, Kubernetes autoscaling).
◆Where it's heading
Windmill is maturing along two tracks at once: a more capable, safer execution runtime (sandboxed multi-tenant containers without a Docker daemon, remote SSH targets) and the enterprise scaffolding around it (environment promotion, distributed tracing, audit export, smarter K8s scale-in). The direction is making the platform safe to run untrusted, multi-tenant code at scale while giving teams real dev/prod discipline.
◆Prediction
Expect continued investment in isolation/runtime breadth and enterprise operability — more sandboxing options, deeper observability, and workspace/promotion tooling. The cadence is fast and consistent, with the daemonless container runtime the most likely lever for new multi-tenant and cloud use cases.

◆Recent moves

  1. 1d ago

    Workspace DuckDB macro libraries

    View source ↗
  2. 1d ago

    Backfill a range of partitions from the asset drawer

    View source ↗
  3. 2d ago

    SCD2 history strategy for DuckLake materialization

    View source ↗
  4. 2d ago

    Local development for data pipelines

    View source ↗
  5. 3d ago

    Dev workspaces paired with a lockable prod workspace

    Adds persistent dev workspaces paired with a lockable prod workspace, where changes reach prod only through manual promotion via Compare & Deploy, with per-environment resources and variables. Brings real environment discipline to the platform, part of the enterprise-maturation track.

    View source ↗
  6. 23d ago

    Kubernetes autoscaling scale-in prefers idle worker pods

    Kubernetes autoscaling now annotates worker pods with pod-deletion-cost so scale-in removes idle pods first and protects busy ones by job age. A best-effort ops refinement that makes autoscaling less disruptive to running work.

    View source ↗