July 2026 monthly report for twentyhq/twenty: implemented configurable health probes for server deployment to increase reliability during startup and operation, and validated across environments; introduced parameterization of startupProbe, livenessProbe, and readinessProbe under server in values.yaml with sensible defaults and a health endpoint path change to /healthz; added schema support; validated with helm lint/template; observed boot-time improvements and avoided CrashLoopBackOff; maintained backward compatibility; prepared for follow-up to cover worker deployment in future.

June 2026: Delivered a feature extension to the Twenty Helm chart enabling pod-level customization for server and worker deployments. Added four pod spec fields (nodeSelector, tolerations, dnsPolicy, dnsConfig) and exposed them under server.* and worker.* in values.yaml, while preserving backward compatibility. The change uses the existing templating pattern to render only when values are provided, ensuring an empty/default output for existing installations. Validated with helm lint and helm template; live validation on a multi-node Kubernetes cluster with a dedicated tainted node and custom DNS (ndots: 1). Server and worker pods scheduled and started successfully. This enables operators to align deployments with node pools, taints, and DNS configurations without forking charts. Co-authored by Charles Bochet; PR #22233, closes #22250.