Production-ready Helm chart with Kubernetes fixes

[khalid244]

Hey @vermorag,

Following up on #119 and your work on the .helm/ chart in PR #294 — I've built and tested a production-ready Helm chart that fixes all the known Kubernetes deployment issues.

Repo: https://github.com/khalid244/cattr-helm
Helm repo: helm repo add cattr-helm https://khalid244.github.io/cattr-helm

What it fixes

• s6-overlay /var/run and timeout issues — emptyDir mounts + S6_CMD_WAIT_FOR_SERVICES_MAXTIME=300000
• MySQL trigger creation fails — --log-bin-trust-function-creators=1 (required for Cattr's migrations that create triggers)
• App starts before DB is ready — init container waits for MySQL TCP
• Cache/session write errors — ephemeral emptyDir mounts for framework cache and sessions
• Migration conflicts — documented workarounds for redmine_instances duplicate table and Sanctum expires_at column

Improvements over current .helm/ chart

Feature	Current .helm/	This chart
Service template	Missing	Included
PVC templates	Missing	Included
Ingress	YAML syntax error	Working, with TLS/cert-manager
APP_KEY	encryptAES (non-standard)	randAlphaNum + b64enc
MySQL	Bitnami subchart	Built-in MySQL (no external deps)
DB startup order	None	Init container waits for MySQL
MySQL triggers	Not handled	--log-bin-trust-function-creators=1
Values style	Custom	Bitnami-style with @param annotations

Tested on

• 3-node Hetzner cluster (MicroK8s)
• Longhorn storage
• nginx-ingress + cert-manager (Let's Encrypt)
• All s6 services start cleanly: nginx, octane, reverb, queue, supercronic
• 144 migrations run successfully
• Full app functional at https endpoint

Happy to submit a PR to replace or improve the existing .helm/ chart if you're interested. Let me know how you'd like to proceed.