v0.8.0

Highlights

Tasks

Declare migrations, smoke checks, and backfills under tasks: next to components. from copies image, env, and profiles from a component. "on": preDeploy and "on": postDeploy run around deploy; "on": manual is CLI-only. deployah run starts any task, including a hook you want to retry.

apiVersion: v1-alpha.5
project: shop
components:
  api:
    image: ghcr.io/acme/shop:1.2.3
    env:
      DATABASE_URL: ${DATABASE_URL}
tasks:
  migrate:
    from: api
    "on": preDeploy
    command: ["migrate", "up"]
  smoke:
    from: api
    "on": postDeploy
    command: ["curl", "-f", "http://api/health"]
  backfill:
    from: api
    "on": manual
    command: ["backfill"]

Quote "on" in YAML 1.1 so it is not read as a boolean.

Scheduled tasks

"on": schedule creates a Kubernetes CronJob in the release. deployah deploy applies it and does not start a Job on that deploy. deployah run still creates a one-shot Job.

tasks:
  cleanup:
    from: api
    "on": schedule
    schedule: "0 3 * * *"
    command: ["cleanup"]

Init wizard

deployah init is interactive and must run from a terminal. It writes deployah.yaml and a platform file. --dry-run prints the spec without saving it.

App manifests move to v1-alpha.5; platform profiles stay platform/v1-alpha.3.

What's Changed

Features

• add tasks for run-to-completion work by @atkrad in #83
• make init a tty-only wizard by @atkrad in #89
• add scheduled tasks as CronJobs by @atkrad in #104

Bug Fixes

• harden helm pending checks and CRD coverage by @atkrad in #82

Full Changelog: v0.7.0...v0.8.0

v0.7.0

Highlights

Stateful workloads

Use kind: stateful for stable network identity (ordinal hostnames, headless DNS) and optional per-pod volumes. Stateless components can take a shared PVC too. Grow disks later with deployah deploy --resize-volumes.

components:
  cache:
    kind: stateful
    image: redis:7-alpine
    port: 6379
    resourcePreset: nano
    environments: [dev]
    replicas: 1
    persistence:
      size: 20Gi
      mountPath: /data

Workers and metrics

role: worker runs long-lived processes with no Service port, Ingress, or ready probe. Enable Prometheus scraping with metrics (ServiceMonitor for services, PodMonitor for workers). Profiles supply metrics.monitorLabels.

components:
  api:
    image: ghcr.io/acme/api:1.0.0
    port: 8080
    environments: [production]
    profiles: [observability]
    metrics: true
  worker:
    role: worker
    image: ghcr.io/acme/worker:1.0.0
    environments: [production]
    profiles: [observability]
    metrics:
      port: 9090
      path: /metrics

# deployah.platform.yaml
apiVersion: platform/v1-alpha.3
profiles:
  observability:
    metrics:
      monitorLabels:
        release: prometheus

App manifests move to v1-alpha.4; platform profiles use platform/v1-alpha.3.

What's Changed

Features

• add kind: stateful workload support by @atkrad in #70
• add worker role and prometheus metrics by @atkrad in #75

Full Changelog: v0.6.0...v0.7.0

v0.6.0

Highlights

CI now runs OSV-Scanner on dependencies and on the linux/amd64 release image, with findings uploaded to Code Scanning. The README also gains an examples/nginx starter and demo GIFs so you can see a local deploy without reading the full docs first.

What's Changed

Features

• add OSV-Scanner CI for deps and images by @atkrad in #61

Documentation

• add nginx example and README demo GIFs by @atkrad in #52

Full Changelog: v0.5.0...v0.6.0

v0.5.0

Highlights

You can now add custom manifests and CRDs under .deployah/ so extras join the Helm release without hand-editing charts. Manifests use a post-renderer; CRDs apply first with --crds create|create-replace.

What's Changed

Features

• add custom manifests and CRDs via .deployah/ by @atkrad in #5

Full Changelog: v0.4.0...v0.5.0

v0.4.0

Full Changelog: v0.3.0...v0.4.0

v0.3.0

Full Changelog: v0.2.0...v0.3.0

v0.2.0

Full Changelog: v0.1.9...v0.2.0

v0.1.9

Full Changelog: v0.1.8...v0.1.9

v0.1.8

Full Changelog: v0.1.7...v0.1.8

v0.1.7

Full Changelog: v0.1.6...v0.1.7