You probably don’t want to; transacting is an intentionally serial process. So any significant work you do inside the transactor will block other transactions from going through.

I see. So maybe my code would do something like transact and afterwards trigger a lambda that makes the slack message?

my aws-fu isn’t great, but that sounds reasonable

https://docs.datomic.com/cloud/transactions/transaction-functions.html#creating

https://github.com/markbastian/replionhas been a great guide for this, although the procedure is much simpler since this summer's new Datomic Cloud.

The new topology makes it easier since there no more is any bastion to go through, and the Datomic instances are publicly exposed to the Internet.

Opening the tunnel is as easy as: ssh -i/<path-to-key>.pem -L 7000: where the X vals are your compute instances public DNS.

7000 assumes that's the port you serve nRepl on.

oh nice thanks a lot

You only need to adjust the security group on the compute instance to let pass the traffic, as shown in Replion's guide (but the guide shows you to do it on the Bastion instance and the compute instance IIRW, but now there is no bastion, only the compute instance to open up).

it's like an "inbound rule" on the security group right?

Yes, let me confirm and screenshot.

which is the security group that counts for a query group stack? There is lambda and loadbalancer group

The one with -NodeSecurityGroup- in the name.

will try making it work 😄

good luck! Things are errily silent when we miss a detail, but otherwise it's quite easy.

what is missing a detail? Ah you mean if something isn't working there is no warnings and such when trying this

Approximately in this order: 1. Network SSH p22: If your addresses are wrong, some error; if your path to the pem key is wrong, quick auth error 2. Network TCP p7000: If you open up the ports on the wrong SG or source CIDER, etc., IIRW connection will fail quickly with some error, or silently after a long timeout 3. nRepl service: If you forget to start your nRepl server on the appropriate port, long timeout IIRW I could be well off in my predictions. I never can remember which response is caused by which detail missed.

haha it works so sick. Can even use cider debug instrumentation on lambdas :pogg:

Nice, Benjamin! 🙂

Nope

yea I fix it differently