postmortems and notes — coming soon.

I'm pulling postmortems forward from the last two years and rewriting the ones worth sharing. They'll land here as I clear them. Pieces in flight:

• · nyc etcd self-removal recovery (rke2, multi-node)
• · emptyDir + postgres: how it actually fails
• · venue provisioning pattern: terraform → ansible → jenkins
• · ~64% aws cost reduction, in three passes

Until then, the FAQ on the home page is the static version of what i think.