MaterializeInc / materialize

The data warehouse for operational workloads.
https://materialize.com
Other
5.68k stars 459 forks source link

[Epic] Cluster-specific `Optimize` pipelines #25273

Open aalexandrov opened 5 months ago

aalexandrov commented 5 months ago

Product outcome

Materialize engineers working on changes in the compute layer that require changes to the Optimize pipelines can perform production experiments on unbilled clusters which force the proposed new version of the pipeline (enabled behind a cluster-specific feature flag).

Discovery

Work items

### Pre-production (discovery)
- [x] #25010
- [x] #25165
### Production (development)
- [ ] https://github.com/MaterializeInc/materialize/pull/25373
- [ ] https://github.com/MaterializeInc/materialize/pull/25544
- [ ] https://github.com/MaterializeInc/materialize/pull/25586
- [ ] https://github.com/MaterializeInc/materialize/pull/25670
- [ ] https://github.com/MaterializeInc/materialize/pull/25811
- [ ] https://github.com/MaterializeInc/materialize/pull/25812
- [ ] https://github.com/MaterializeInc/materialize/pull/25628
- [ ] https://github.com/MaterializeInc/materialize/pull/26460
- [ ] Add support for `EXPLAIN SUBSCRIBE` (stretch goal)
- [ ] Add support for `EXPLAIN COPY TO` (stretch goal)
### Post-production (rollout)
- [x] Merge design doc that reflects the merged version of the code
- [x] Use it to run `enable_eager_delta_joins` experiments
- [x] Use it to run `enable_variadic_outer_joins` experiments

Decision log

ggevay commented 2 months ago

This really seems done.