Effective Datomic Imports

Transcript

1. Effective Datomic Imports
2. Ryan Neufeld @rkneufeld
3. Ryan Neufeld @rkneufeld
4. Effective Datomic Imports
5. Staples Innovation Labs
6. Big Reporting
7. Trouble Out
8. Trouble Out Trouble In
9. Stop.
10. Craig Time.
12. (def real-tx d/transact) ! (with-redefs [d/transact (fn [conn tx-data] (println tx-data) @(real-tx conn tx-data))] ;; A bunch of business logic ) Gross Hacks
13. Tip: Separate Tx Data from Txing
14. Pipeline
15. Pipeline Shape
16. Real-life Transactor
17. DynamoDB
18. Did I use 500 write units… or six?
19. Throttled
20. Free & Clear
21. Take-aways • Separate Tx from Tx’ing • Batch to 100 datoms/tx • Pipeline txes asynchronously • Use realistic backend/ settings • Use CloudWatch to watch StorageBackoff and AlarmBackPressure (carefully!) • Capacity Planning Docs
22. Staples is hiring! http:/ /careers.staples.com/sanmateo/ • Clojure Engineers • Data Scientists • Product Managers • Quality Assurance • DevOps Engineers • UI/UX Designer
23. So is Cognitect! http:/ /cognitect.com/jobs • Datomic Support Engineer • Clojure Developers (Contract)
24. Questions? Ryan Neufeld @rkneufeld rkn.io