mistercrunch commented on PR #33837:
URL: https://github.com/apache/superset/pull/33837#issuecomment-3146705169

   > Having an officially supported chart is a big asset to the project
   
   Totally agree. Ideally people using the chart actively can help maintain it. 
One thing I've found in devops is that trying to write a complex, 
multi-services abstraction (pick you own database, pick your own caching layer, 
pick your own message queue, ...), with hundreds of config keys is extremely 
challenging. Like the abstraction itself (say the helm chart of k8s operator) 
ends up being much more code than any given implementation would need.
   
   The main dillema around the helm chart here as a contributor put it, it's to 
decide whether it's just "a reference implementation that works well and is 
bulletproof" VS "an abstraction that allows you to configure your system in all 
sorts of ways". I found the former would be much easier for us to maintain and 
provide guarantees around.
   
   About the config system, I'm working on something that will make Superset 
much more configurable, where every config key that's static (serializable to 
json, not a python object) will be available through env vars (prefixed by 
`SUPERSET__`) or will support using the metadata database for storing them 
(defined in a `DATABASE_MANAGED_CONFIG_KEYS = []` or similar), which will be 
fully dynamic/runtime. Should help quite a bit.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to