hey everyone, I had a clojure app using datomic(dynamo) deployed to aws. The app was working before but now I started getting errors like this

17-11-08 06:42:15 221c15fc989f ERROR [io.montaigne.api.server:31] - datomic connection error clojure.lang.ExceptionInfo: Error communicating with HOST 10.238.58.21 or ALT_HOST 54.190.108.228 on PORT 4334 {:alt-host "54.190.108.228", :peer-version 2, :password "EDITED, :username "EDITED", :port 4334, :host "10.238.58.21", :version "0.9.5554", :timestamp 1506450187882, :encrypt-channel true}
17-11-08 06:42:15 221c15fc989f ERROR [io.montaigne.api.server:32] - datomic connection error details Error communicating with HOST 10.238.58.21 or ALT_HOST 54.190.108.228 on PORT 4334

hi - I'm getting a 'Transactor not available error' when doing a large data import, the full stack here: https://gist.github.com/jonpither/17a90989d42569988db30d2171e5d58e. Is this recoverable on the Peer process? The transactor seems to be able to carry on with heartbeats etc.

I've put a retry code + exponential-backoff around the transact-async call - wondering if other's have taken this approach?

Sending many txs in a tight loop without back pressure (deref) or sending txs that are really big can both cause that

The tx may succeed, but the transactor is too overwhelmed to heartbeat

Hi, I have an entity that consists of a unique attribute (squuid), a third party id (unique identity) and a bunch of other attributes. What is the best way to upsert data based on third party id, without overriding my squiid if it already exists? i.e. I want to generate a squuid the first time but not on the subsequent updates. This seems to be a common use case so I wonder if a custom database function is reasonable