You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be great if RudderStack Server supported configuration by means of text files. In principal, this could enable the following workflow:
actively developing a Server configuration in a dev or test environment
versioning this in source control
ultimately, deploying to a production environment when deemed ready
In my team, we follow this sort of pattern already with Kubernetes using Helm, and have found that the declarative nature of Helm charts gives us confidence in its robustness while allowing changes to be tracked in a way which is comfortable for developers (namely, in Git repos). Although I appreciate that RudderStack Server is a very different tool to Kubernetes, I think that this sort of workflow could be a compelling one for companies invested in CI, CD and dev-ops in general.
I can see challenges relating to environment secrets, the various resources (e.g. sources and destinations) changing rapidly as RudderStack evolves, and the potential complexity of a scriptable templating language to allow for this to work, but perhaps a basic implementation could allow for a solution to be developed iteratively.
Note that I was asked to raise this after having had a Slack conversation with Pradeep (I think).
The text was updated successfully, but these errors were encountered:
It would be great if RudderStack Server supported configuration by means of text files. In principal, this could enable the following workflow:
In my team, we follow this sort of pattern already with Kubernetes using Helm, and have found that the declarative nature of Helm charts gives us confidence in its robustness while allowing changes to be tracked in a way which is comfortable for developers (namely, in Git repos). Although I appreciate that RudderStack Server is a very different tool to Kubernetes, I think that this sort of workflow could be a compelling one for companies invested in CI, CD and dev-ops in general.
I can see challenges relating to environment secrets, the various resources (e.g. sources and destinations) changing rapidly as RudderStack evolves, and the potential complexity of a scriptable templating language to allow for this to work, but perhaps a basic implementation could allow for a solution to be developed iteratively.
Note that I was asked to raise this after having had a Slack conversation with Pradeep (I think).
The text was updated successfully, but these errors were encountered: