Deployment, Development, and Version Control

[guardhunt (Christopher Hunt)]

I am relatively new to elastic and am looking for some input on choosing a deployment, ongoing development, and version control strategy for my Elsatic Stack. I am close to building a production version of elastic am looking to get some input on how other people typically manage their stack after its in production. Does anyone use version control, like git? IF so, what do you include, what do you exclude and why? Do you have a mirrored development stack that you test your changes on before launching to to production, like you would in any other development strategy? Does this need to be on an identical system?

Any thoughts or answers are appreciated, thanks!

[mujtabahussain]

@guardhunt

manage their stack after its in production

When you say manage what exactly do you mean? Are you asking in terms of managing code that builds/deploys/tests the ES stack or something w.r.t. data that is in ES itself?

[guardhunt (Christopher Hunt)]

Both really. I am interested in seeing what people do and how they do it both internally and externally. In my application I am dealing with time-based log data so I am coming up with a strategy to handle that, and then handle long term management for builds down the road.

[mujtabahussain]

Most of my stack is deployed on AWS myself. So the entire orchestration is done via CloudFormation templates. These are version controlled.

All the elasticsearch data is backed up by snapshotting to S3 and the scripts to do that are cron enabled and also version controlled.

All of the deployment is done via a CI.