Yep, same here. @U0AQ1R7FG

oh sorry, thanks for bringing that to my attention

hm DNS is pointing to the correct address now but the wrong app is serving responses. i’ll need to look into this more

Maybe just DNS taking a while to propagate?

Hooray! I bet folks would be interested to hear what the internal advocacy process was like.

Congrats (touch 🪵 )! What is the current stack?

This is limited to a group that I’m involved with, and the system will be running independently of other groups in the organization. Until now, the only languages in use have been C# and Python. However, I need to run ETL processes in an AWS Lambda that reads files from S3 and generates TTL files that go back to S3. That makes it a black box from most of the company’s perspective. Also, some of the necessary libraries for some of the input files are only available in Java, which meant I needed the JVM anyway.

Ooooh! ETL and core.async! Perfect together! Have fun,

Initially, I wrote it to run on my local machine, and that was fine. Only a few people knew I’d used Clojure. But when we moved it to a CI pipeline we needed the uberjar build tooling, and that’s where it got exposed to others (since the people who administer the pipeline didn’t have Clojure available on those systems). Fortunately, the 3 devs working on this all know Clojure, and I’m senior enough that I was allowed to propose it for trial (we skipped “Assess” because I’d already done that) 🙂

No core.async

I did a hairy ETL going from a Postgres instance thru my ETL/core.async MacBook app then out to a third aggregator then on to a ElasticSearch...hang on...https://tilton.medium.com/backpressure-99501f23881f. The only time I heard my fans come on!

In this case, the entire dataset can fit in RAM, so long as you expand your default heap

The largest one is RxNorm, which is 1.2G of raw text

Sweet. We were doing geodata. Canada's coastline alone was 140mb. Norway was close!