NewsLab
Jun 28 17:23 UTC

Show HN: DBOSify – Drop-in Temporal replacement built on Postgres (github.com)

87 points|by KraftyOne||19 comments|Read full story on github.com

Comments (19)

19 shown
  1. 1. KraftyOne||context
    Hi HN, I'm Peter, creator of DBOSify. Here to answer any questions you have!
  2. 2. actionfromafar||context
    What would it involve to get DBOSify for C#?
  3. 3. KraftyOne||context
    We'd need DBOS in C# first! It's something we're considering for the future--if you're working in C#, happy to chat about it, reach out at peter.kraft@dbos.dev.
  4. 4. hatefulheart||context
    You can add a dev dependency in your pyprojec.toml with uv, look it up.
  5. 5. raykyri||context
    Hi! Do you have an example that includes a full agent workflow?
  6. 6. KraftyOne||context
    Take any Temporal example and swap the imports and connection strings, it's a drop-in replacement :)
  7. 7. x3ro||context
    Hey Peter. I know this is super unrelated, but I’ve used the contact form and sent an email in order to understand what the hell a conductor license costs for self-hosting. So far i’ve only gotten automated requests for feedback as a response.. That’s certainly not encouraged me to dig deeper into using dbos.
  8. 8. KraftyOne||context
    I'm very sorry about this. I'll figure out why the form wasn't responded to. In the meantime, email me directly at peter.kraft@dbos.dev.
  9. 9. secondrow||context
    Hi x3ro - dbos marketing person here. Sorry about the contact form, which was broken for a few days last week following changes we made to the cookie management on the website.
  10. 10. kwkelly||context
    Is it compatible with Aurora DSQL? Or does it rely on Postgres features they do not implement?
  11. 11. KraftyOne||context
    Unfortunately, DSQL is still lacking many core Postgres features (most notably foreign keys) that DBOS and DBOSify depend on. DBOS works with most flavors of Postgres though, including regular Aurora/RDS, Cloud SQL, AlloyDB, Supabase, Neon, CockroachDB, and others.
  12. 12. unsnap_biceps||context
    I'm a little unclear on this. You say it's a replacement for Temporal, but the GitHub repo says it's a replacement for the temporal python client SDK?

    Then in the description you say "This lets you run durable workflows, activities, signals, updates, retries, and recovery without needing any infrastructure except Postgres." but your diagram shows worker nodes outside of the Postgres server, so you do need infrastructure beyond the Postgres instance?

  13. 13. lmz||context
    I think "without needing any infrastructure" is in contrast to Temporal, which has its own server in front of the storage.
  14. 14. lelandbatey||context
    I asked the DBOS folks about this before; the idea is that there is no "coordination node", only the workers and the DB. See DBOS folks previously: https://news.ycombinator.com/item?id=45186494
  15. 15. KraftyOne||context
    Exactly, Postgres as an orchestrator instead of a Temporal server as an orchestrator. This page goes into more detail (for DBOS, but DBOSify is the same principle but using the Temporal API directly): https://docs.dbos.dev/explanations/comparing-temporal
  16. 16. swyx||context
    (ex temporal employee) believe it or not this is SIMPLER than temporal's actual workers. what you're seeing is a "DBOSify worker" embedded in each application server, which is not the same thing as temporal's conception of workers (which are separate from the app server). i havent spent much time with dbosify but i'd say this is closer to a "second client" than a full worker... just a terminology issue
  17. 17. antman||context
    Can DBOSfly pass the temporal python SDK test suite? Are there any features mot yet implemented?
  18. 18. KraftyOne||context
    Yes it can (with documented exceptions)! The suite is here: https://github.com/dbos-inc/dbosify-py/tree/main/tests/confo...

    Here's the documentation: https://github.com/dbos-inc/dbosify-py/blob/main/docs/ARCHIT...

  19. 19. aliasxneo||context
    I recently needed an OSS durable execution engine and chose River (I use Go) over DBOS because I wasn’t convinced I wouldn’t get rug pulled by a VC backed company. Both are technically commercial, but River seemed to be doing a lot more OSS friendly moves lately than DBOS, and AFAIK they are self bootstrapped.