FHIR Chat · Recommended Performance Metrics · bulk data

Stream: bulk data

Topic: Recommended Performance Metrics


view this post on Zulip James Bradley (Jul 30 2020 at 19:44):

We are currently investigating the technical requirements for supporting bulk data export. Since the IG doesn't provide recommended performance metrics, I was curious if there are performance metrics that other implementors are working towards and what are those metrics? For example: "Given X number of resources being returned we are shooting for it to take N hours."

It would be interesting to see if there is some general consensus around what an acceptable amount of time is to return a Bulk request.

view this post on Zulip Dan Gottlieb (Aug 05 2020 at 12:49):

@Michael Hansen @Lee Surprenant @Richard Braman - did you address this in your implementations?

view this post on Zulip Lee Surprenant (Aug 05 2020 at 16:43):

I think its really use-case dependent. To date, we've heard more stringent requirements on the import of data rather than its export. With that said, having some ballpark numbers can be helpful; these were the numbers we collected for our server the last time we tested it: https://github.com/IBM/FHIR/issues/1128

view this post on Zulip Paul Church (Aug 05 2020 at 16:56):

+1 to Lee, import performance is "never enough" but export gets virtually no requirements. I think any export in the thousands of resources per second should be enough for most cases.


Last updated: Apr 12 2022 at 19:14 UTC