@sharkinsspatial, @hrodmn, and I were discussing today the load that titiler-cmr and titiler-multidim place on NASA's S3 credential providers. We discussed building some lightweight infrastructure for caching temporary credentials across all the VEDA lambdas. This would reduce cost (not dramatically though) and reduce risks of causing outages due to unmanageable load.
Before going down this path, we wanted to check with @anayeaye if data services has already done any work on this.
We also realize that this is not essential for any near-term deliverables, and shouldn't be prioritized over that work. Still, we hope that we could sketch something relatively quickly and in the nearish future.
@sharkinsspatial, @hrodmn, and I were discussing today the load that titiler-cmr and titiler-multidim place on NASA's S3 credential providers. We discussed building some lightweight infrastructure for caching temporary credentials across all the VEDA lambdas. This would reduce cost (not dramatically though) and reduce risks of causing outages due to unmanageable load.
Before going down this path, we wanted to check with @anayeaye if data services has already done any work on this.
We also realize that this is not essential for any near-term deliverables, and shouldn't be prioritized over that work. Still, we hope that we could sketch something relatively quickly and in the nearish future.