I've been running both Flux and Make for different teams in our community for about a year now, primarily for data integration and internal tooling. One area where they've really diverged in practice is error handling and recovery, which can make or break a production workflow.
Flux, with its focus on data pipelines, treats errors as a first-class citizen. Its "Dead Letter Queue" pattern for failed rows is something we've come to rely on. If a transformation fails on a specific record, that record is quarantined with a detailed error context, and the rest of the pipeline keeps running. The built-in retry logic is also very granular—you can set policies based on error type. The downside? Configuring this properly has a learning curve, and for simple, all-or-nothing tasks, it can feel a bit heavy.
Make, on the other hand, feels more oriented towards the automation of app-to-app workflows. Its error handling is visual and immediate in the scenario map. A failed module turns red, and you can set up a dedicated error-handling route right there in the same scenario, which is great for quick fixes and notifications. However, I've found it less robust for partial failures. If one item in an array operation fails, it often fails the whole module unless you build a lot of extra logic around it.
I'm curious about others' lived experience. For those of you who've pushed these platforms beyond simple "happy path" automations:
* Which platform's error handling has caused you fewer headaches in production?
* Have you found one easier to debug when something goes wrong at 2 AM?
* Any specific features or workarounds for partial failures in Make, or for simplifying the setup in Flux?
Looking to gather some real-world pros and cons to share with the team.
— Eric
Keep it civil, keep it real.