@calumabarnett has requested changes with the ring assignments:
I think I would frame hold solely around technologies we currently use (and are happy to continue using for the time being) but don’t want to adopt in new projects
If we think something is not worth investigating, I don’t think it should be on the radar
Retire is where we should be taking steps to replace
This ticket is about confirming what ring assignments we should use.
A good example to consider is black: we should be replacing with ruff but perhaps not worth the hassle if everything is working ok.
Another example is dagster: there's too much overlap with Airflow so we it's not worth investing time prototyping without first agreeing with AP
Note that we want to limit the number of blips to 25 per category so altogether 100 to make it meaningful and maintainable, so the assignments should help us to limit the number of blips on the radar.
Options
Option 0: Zalando
Assignment
Reaction
Short Explanation
Example
Radar
ADOPT
🚀
Expand / carry-on using
Inner
TRIAL
👀
Start using
Third
ASSESS
👍
Prototype
Second
HOLD
☹️
Not (yet) recommended, migrate if possible
Outer
Option 1: Existing
Combine TRIAL into ADOPT, and split off RETIRE from HOLD
Assignment
Reaction
Short Explanation
Example
Radar
ADOPT
🚀
Start / expand / carry-on using
create-a-pipeline, AWS, Python, tech-radar
Inner
ASSESS
👍
Prototype
dlt
Second
HOLD
👀
Not (yet) recommended
dagster , sql mesh
Third
RETIRE
☹️
Migrate to alternative if possible
glue 2.0, glue, register-my-data, pulumi
Outer
CLOSE
👎
Not relevant
typescript
Option 2: Calum's suggestion
Exclude technologies which are not (yet) recommended and split up RETIRE into REPLACE and RETAIN
Context
@calumabarnett has requested changes with the ring assignments:
This ticket is about confirming what ring assignments we should use.
A good example to consider is black: we should be replacing with ruff but perhaps not worth the hassle if everything is working ok.
Another example is dagster: there's too much overlap with Airflow so we it's not worth investing time prototyping without first agreeing with AP
Note that we want to limit the number of blips to 25 per category so altogether 100 to make it meaningful and maintainable, so the assignments should help us to limit the number of blips on the radar.
Options
Option 0: Zalando
Option 1: Existing
Combine TRIAL into ADOPT, and split off RETIRE from HOLD
Option 2: Calum's suggestion
Exclude technologies which are not (yet) recommended and split up RETIRE into REPLACE and RETAIN
Option 3
Same as option 2 but split off PAUSE from CLOSE