Helm

The Helm chart is the recommended way to run PlotPress on Kubernetes. It expects:

• A PersistentVolumeClaim for /var/lib/plotpress (SQLite).
• A ConfigMap or git-sync sidecar (or both) supplying the workspace at /workspace.
• A Secret for environment variables referenced by ${VAR} in connections.yaml and plotpress.yaml.

Add the repo

helm repo add plotpress https://octopyth.github.io/plotpress
helm repo update

Minimal values

# values.yaml
image:
  tag: v0.1.0

ingress:
  enabled: true
  className: nginx
  hosts:
    - host: plotpress.example.com
      paths: ["/"]
  tls:
    - secretName: plotpress-tls
      hosts: [plotpress.example.com]

persistence:
  enabled: true
  size: 5Gi
  storageClass: hcloud-volumes

# Workspace can come from a ConfigMap or a git-sync init container.
workspace:
  source: gitSync
  gitSync:
    repo: https://github.com/example/plotpress-workspace
    branch: main
    interval: 60s

env:
  WAREHOUSE_PASSWORD:
    secretRef:
      name: plotpress-secrets
      key: warehouse-password
  OIDC_CLIENT_SECRET:
    secretRef:
      name: plotpress-secrets
      key: oidc-client-secret

Install

helm install plotpress plotpress/plotpress -n plotpress --create-namespace -f values.yaml

Workspace strategies

Strategy	When to use
gitSync	Workspace lives in its own git repo; init container clones and syncs every interval. The default.
configMap	Small workspaces with no secrets in dashboards. Reload requires a rollout.
pvc	You manage the workspace out-of-band (e.g. mounted from an existing PVC).

Upgrades

The chart runs PlotPress as a Deployment with replicas: 1. SQLite forces single-writer; multi-replica needs Postgres (planned). For now, upgrades are short rolling-restarts (~1s of unavailability).

Flux integration

A HelmRelease example for Flux:

apiVersion: helm.toolkit.fluxcd.io/v2
kind: HelmRelease
metadata:
  name: plotpress
  namespace: plotpress
spec:
  interval: 10m
  chart:
    spec:
      chart: plotpress
      version: ">=0.1.0 <1.0.0"
      sourceRef:
        kind: HelmRepository
        name: plotpress
        namespace: flux-system
  valuesFrom:
    - kind: ConfigMap
      name: plotpress-values
    - kind: Secret
      name: plotpress-values-secret

Resources

The Go backend is small. Reasonable starting point:

resources:
  requests:
    cpu: 50m
    memory: 64Mi
  limits:
    memory: 256Mi

Most of the cost is the connection pools to your data sources; tune those in connections.yaml rather than at the pod level.