Enhance production coverage in the documentation

A common request from customers about the Nuxeo production architecture is about what procedure should one plan if a component upon which Nuxeo relies misbehaves?

• E.g. How to best configure ES/DB/Redis for failover/failback/high availability?
• E.g. If ES/DB/Redis collapse, what to do in what order? Ia there a risk of data loss, and in what circumstances?
• E.g. When to shutdown and restart Nuxeo in case a misbehavior is noticed?
• E.g. Is it necessary to shutdown Nuxeo in case some production event is noticed? (e.g. re-registration)
• E.g. To help further diagnose a production issue, what elements to gather in first intention to help investigations and avoid need for next occurrence to happen to gather the correct information? How to get and gather these pieces of information?
• E.g. What to monitor to assert Nuxeo application health?