How to use retry mechanism whenever any data failure happens from an userevent script to external system

That's a broad question and quite non-specific. There could be many approaches. Basically you'd have to capture the failure and retry keeping track of how many attempts you've made until you've hit a "reasonable" limit. You'll likely hit some kind of time out though I imagine which you cannot do much about.

often a better solution than retry in a UE script is to add a scheduled/MR script to back up the UE. If you don't strictly need 'near real time' flow of data you can discard the UE entirely and stick with the scheduled/MR script alone.