LinkedIn Carousel: Error Handling at 2am

Post type: Technical · 10 slides
Blog post: https://www.nuvikatech.com/blog/posts/error-handling-auto-file


SLIDE 1 — Cover

Headline: What happens when your auto-filing system fails at 2am and a customer’s breach record is mid-resolution?

Sub-line: The error handling design decision that changed how we think about side effects in production systems.


SLIDE 2

Label: THE NAIVE ANSWER

Headline: Let the exception propagate. Roll back the transaction.

Body: The breach record stays unresolved. The customer gets an error. Someone investigates. This is wrong — and the reason reveals something important about system design.


SLIDE 3

Label: THE KEY INSIGHT

Headline: Breach resolution and claim filing are not the same operation.

Body: Breach resolution: a deterministic database write. Claim filing: an external API call to AWS, Azure, or GCP. Different reliability profiles. Tying them together lets a vendor support tier issue corrupt a breach record.


SLIDE 4

Label: THE SOLUTION

Headline: Total error containment. The filing step can fail — the breach resolution cannot.

Body: The auto-file code is wrapped in a try/except that catches everything and never re-raises. The breach is always left in a defined, actionable state regardless of what the provider API returns.


SLIDE 5

Label: THE STATE ON FAILURE

Headline: Any exception → breach.status = "Assisted Filing Required". Never stuck. Always defined.

Body: “Assisted Filing Required” is not a failure state — it’s a defined outcome. Fintropy generates a pre-filled support case with evidence attached, filing deadline displayed, and direct link to the provider portal.


SLIDE 6

Label: THE INNER TRY/EXCEPT

Headline: Even the alert creation is wrapped. An alert failure must never mask the breach status.

Body: If the alert system is down, we log and move on. The breach record is intact. The logging exists so the filing failure doesn’t go unnoticed — but it cannot block the primary operation.


SLIDE 7

Label: THE TRADEOFF

Headline: Total error containment makes silent failures possible.

Body: If the auto-file system is broken, every breach gets “Assisted Filing Required” without anyone noticing. We mitigate with: log monitoring on Auto-file failed, ratio metrics on Filed vs Assisted, integration tests before every release.


SLIDE 8

Label: THE ASSISTED PATH

Headline: Assisted filing isn’t degraded — it’s a well-designed fallback.

Body: Pre-filled case text, evidence attached, provider portal link, deadline countdown. A customer can file manually in under 5 minutes. Dramatically better than starting from scratch.


SLIDE 9

Label: THE PRINCIPLE

Headline: Side effects must never break primary operations. If they fail, they should fail gracefully.

Body: Filing is a side effect of resolving. Resolving is the primary operation. Design accordingly. This principle applies to any system where a secondary action can fail for reasons outside your control.


SLIDE 10 — CTA

Headline: Want the full story?

Body: The full error handling design: the state machine, the assisted filing path, and the tradeoff between correctness and silent failures.

Link: nuvikatech.com/blog/posts/error-handling-auto-file