New sales demo / restructure existing demos

Non sono indicati file, test o punti di ingresso. Inizia facendo un inventario delle demo esistenti e confrontando i loro confini di distribuzione e manutenzione con le integrazioni e le funzionalità trasversali richieste. Per considerare completato il lavoro, è necessario concordare una struttura delle demo, un modello di responsabilità e un ambito concreto prima di poter iniziare l’implementazione.

Current Situation and Problem Statement
Coming from a sales point of view, the current list of demos we have is not ideal for my specific use case, which I know is very non-normal :)

For me the normal case is, that I have to pick a demo and spin it up before a meeting and then show it to some people who are interested in - I don't know what they are intersted in before the meeting :) At most I know a few products they want - and unless we score a lucky hit there is usually not a demo covering all of them.

On top of that, there are additional topics that may or may not come up during the discussions:

• GitOps
• Security
• Logging
• Monitoring
• Integration with tool xyz (see below for more details on this)

At the moment we have specific demos for some of these topics, but that usually means having a second demo deployed if I want to show any of these topics. Ideally these cross-cutting functionalities could be enabled on all our demos - at the cost of additional complexity and overhead.

Then there is a long list of tools that are often brought up in discussions, which in theory integrate nicely with our stack, and it would be great being able to "just" show this ad-hoc:

• OpenMetadata
• dbt + Astronomer Cosmos
• Lakekeeper
• Polaris
• GarageFS / ...
• DataHub
• dagster
• ...

Sure, not all of those are super-relevant, but for example having just OpenMetadata, dbt and Lakekeeper in a demo would probably cover 80% of what I am currently being asked about.

For me, and again, super subjective and just my personal situation, this leads to fairly often having to just say "this is possible" or "i can show you later" in these demo meetings, when in principle we'd have everything in place, because I "guessed" wrong, when standing up a demo.
For me personally, the ideal situation would just be one demo that does it all:

• logging
• metrics
• security enabled
• audit logs ingested
• some random use case or two
• Trino set up with multiple backends
  • Postgres
  • OpenSearch
  • Kafka
  • S3
  • Hdfs
  • Lakekeeper
  • Hive
• Airflow running a dbt dag with Cosmos
• OpenMetadata ingesting tables and lineage
• Keycloak
• Policies
• whatever else

That way I could spin up the same demo every time and know I have what I need. But I know that I am not the norm there, so this is not an ask for that demo, I'd just discuss possible ways of improving the situation for me. And if there is nothing to do, that is fine too, it has worked so far.

Ideas

• Merge a few demos into this one big demo, might actually cut down on the net effort needed to maintain demos, not sure
• Maintain the demo as a sales owned but not officially supported demo, no testing for releases, we are in charge of maintaining it and extending it as needed -> totally fine, I am happy owning it, but that'll take time and is not my job, I cannot decide that on my own
• many more options