Data and secrets: a replicated relational database with off-site backups and weekly automated restore drills ; a dedicated secrets manager as the single source of truth, delivering credentials to workloads and CI through reviewed, automated channels. Application repositories never hold vault or cluster credentials.
Observability and delivery: full metrics, logging, and tracing coverage; CI pipelines that render and diff every infrastructure change for review; automated dependency updates; and ISO 27001-scoped change management enforced in the repository itself.
Production reliability, observability, and incident response for AI systems operating in a regulated environment, on the platform described above (healthy and actively invested in, not a rescue project).
...