For the time of the demo, we should definitely bump up resources, and we should test bumping up and its effects on performance before the demo.
Tasks:
@batpad to find a time with @zacharyDez and @geohacker where we can temporarily bump up resources, and then test fetching data from the jupyter notebooks and QGIS, and see if we have acceptable levels of performance for the demo
@batpad to work on some tuning of the resources to figure out what values look good for the demo, and then bump up resources one day before demo and remember to turn back down after
@bitner it might be nice to chat about the database setup if you have a bit? We're currently running the postgres database in the cluster - if there's simple things we could do to speed it up, would be great - of course, we're quite conservative with resource allocations right now, so we can also bump those up.
We really just need this to perform at acceptable speed for the demo - I ideally don't want to get into either extensive db tuning or setting up caching layers - we know we can do all that in an actual production environment if we need to.
Fetching data from the eoAPI instance is currently quite slow: https://eoapi.ifrc-risk.k8s.labs.ds.io/
This is mostly because we have been extremely conservative with resource allocations, both for the database as well as the application containers: https://github.com/developmentseed/eoapi-risk/blob/main/deploy/helm/eoapi/values.yaml#L8
For the time of the demo, we should definitely bump up resources, and we should test bumping up and its effects on performance before the demo.
Tasks:
We really just need this to perform at acceptable speed for the demo - I ideally don't want to get into either extensive db tuning or setting up caching layers - we know we can do all that in an actual production environment if we need to.
cc for visibility: @sunu @Rub21