On: We saved $50k/year with a Go microservice coded in a hackathon

[hnav]

Elasticsearch works for this use-case quite well, you'd store a fairly straightforward representation of the MySQL row as a document, query by the fields you're interested in and ask for aggregations on the matching documents. Common bitsets get cached automatically.

[strukturedkaos → hnav]

This exactly how we implemented the rules engine in Kevy. We construct an Elasticsearch query based on the rules selected in the UI. Then we use the scroll API to retrieve the matching documents.

[EpicEng → hnav]

Or maybe just write a less insane SQL query to begin with?

[baud147258 → EpicEng]

If I remember right, elasticsearch query are pretty insane to begin with, so it's another problem.