Let’s say I have the following:
[A]
|
|
|
[B]
|\
| \
| \
[C] [E]
| |
| |
| |
[D] [F]
A is a list of strings
B fetches data from an API using the given string
C performs some kind of work on the data
D writes to a db
E logs errors to a webservice
F posts a message to slack if there are a lot of errors in a small time window
Is it possible to handle errors like with Flow? As in, B can output 2 types of events: data and errors. Data is consumed by C and errors are consumed by E. In GenStage I know I can use a GenStage.PartitionDispatcher with a hash function that assigns events to one partition or the other depending on if they are regular data or errors. But I’m wondering if this is possible within the Flow API?
The best I can come up with is that B would send messages to a producer stage on error, which would be the start of a separate Flow (that consists of E and F). Is there a cleaner way?
