DevOps for "11 Nines" Oracle Availability

[alo9507]

One of the major reasons I became interested in ChainLink over every other blockchain endeavor is because of its focus on smart contracts over tokens, and the way in which operating nodes and writing external adapters allowed me to leverage my traditional software development skills in this new space.

In particular, since availability is a major point of reputation and trust for node operators, we'll want our infra to be rock solid, aka "11 nines" available (99.999999999%).

Let's use this discussion to organize ideas on how to make that possible with solid DevOps, Infrastructure as Code, orchestration and cloud native practices.

Linkpool is doing well in this I've noticed, publishing Terraform scripts for operating ChainLink nodes in the cloud: https://registry.terraform.io/providers/linkpoolio/chainlink/latest/docs

[mktcode]

I'm reading to catch up on these topics:

https://www.weave.works/technologies/going-cloud-native-6-essential-things-you-need-to-know/
https://stackify.com/what-is-infrastructure-as-code-how-it-works-best-practices-tutorials/
https://www.gitops.tech/

This resonates with me a lot.

[alo9507]

Absolutely. DevOps rightfully takes a back seat to application development until there are people who aren't core team members working on and using the platform. But when the time comes to deploy...it comes all at once and often with a surprisingly large effort overhead.

But we'll be prepared!

[alo9507]

Post-launch idea: Publishing our cluster health checks live on the website. Particularly useful if the GitHub and Twitter oracle network is white-labelled for other projects to use as well.