Fork me on GitHub
#datomic
<
2021-08-02
>
joshkh13:08:56

did the 884-9095 upgrade happen to somehow override the /ping REST resource for a deployed HTTP ion? prior to the upgrade my API had a /ping route with a custom response. after the upgrade i just get back Healthy Connection

Daniel Jomphe14:08:40

Seems so! Here, our custom ping resource is behing an e.g. /some-path/custom-ping URL so I didn't remark the potential change, but I now see that /ping responds just like you do.

joshkh18:08:40

thanks for confirming @U0514DPR7. i'm glad i wasn't using our custom response for anything important 😉

🙂 3
redinger20:08:01

That’s interesting that this wasn’t an issue until after you upgraded to the new release. AWS documents the /ping path as reserved for API Gateway for their own health checks. https://docs.aws.amazon.com/apigateway/latest/developerguide/api-gateway-known-issues.html (Although, they don’t mention the HTTP API in that page, it is also the case they use that path there as well)

enn14:08:15

I’m curious how other Datomic users handle the need for aggregating counts of entities efficiently. I’m thinking of things like “I want to know the number of albums in every category”, where the numbers of both albums and categories are large and can grow without bound. Do you just do aggregation and count from scratch every time you need to report this info? Do you store it? If you store it, do you use something like transaction functions to keep it up-to-date? We are finding that where there are many possible state transitions it can become costly in engineering time to ensure that these aggregate values are always updated correctly.

xlfe00:08:36

We're using either the datoms api or the index api which are very quick. The key for us has been having the right attributes indexed from the start and using the low level accesses