Each teammate knows what to do if another teamate is absent, including who handles what, how, when, and why.
When there's an issue, then we have a way to triage it, handle it, learn from it, and improve because of it.
When we have dangerous setups, we use lockout/tagout.
Each of our significant projects has a risk log, and its kept up to date.
We use backups that match our planned level of risk, and we regularly verify that the backups work.