Sort of a laundry list but if you hit most of these you're in pretty good shape.

Cassandra snapshot backups

Introductory operational information about Cassandra.

Operations history lesson.

Things to consider when optimizing at the 99th percentile. Variance is key and is one of the biggest reasons EBS blows.

Counterintuitively, he suggests synchronizing variances across a cluster (for example, a Puppet run). Even though for that moment in time all requests are slow, it's better to do that than let every request potentially be slow because it fanned-out to a temporarily slow node.

A thousand times, "yes."

(Dumping GitHub watches into Pinboard.)

AWS node discovery tools.

Interesting CM-focused analogue to Chubby/ZooKeeper/Doozer. I can't help but think the high-level API has muddled the design. I'd love to see that same API implemented in terms of Doozer.

Some good, if a bit obvious, advice on operable cloud infrastructure. "Manage your systems from outside the cloud," is one I'm not so sure about, as it implies a more permeable firewall and/or more directly Internet-accessible system components, which may not be an option for many companies because of, among other things, PCI-DSS standards.

Plugs for Puppet are always good, too.

Ops automation