Field notes

Notes on jobs that report success and don't do the work.

Silent failure in regulated filing, scheduled data obligations and long-lived .NET and Azure estates. Why monitoring misses it, and what actually catches it.

6 notes · newest first

Latest6 July 2026

How to design an n8n workflow that degrades gracefully instead of failing silently

Most n8n workflows are built to succeed. This post introduces explicit degradation tiers so your automations handle partial failure without going dark.

ai-workflow-ops
n8n
workflow-design

Read the note

Earlier notes

29 Jun 2026
The Make scenario that runs every night and has never once been checked
Most Make automations reach production without alerting, log review cadence, or a named owner. 'It ran' is not the same as 'it worked.' Here is how to build observability in deliberately.
Read
22 Jun 2026
What silent data corruption actually looks like in a production n8n workflow
The failure mode most workflow builders miss is not the loud crash. It is the field that maps wrong, the payload that truncates, the LLM node that returns a plausible but incorrect value. This post documents the patterns and the design constraints that prevent them.
Read
22 Jun 2026
n8n error handling that works in testing silently fails in production
Most n8n error handling is tested against predictable failures. Production breaks differently. Here is why your error nodes stay quiet while bad data flows downstream.
Read
10 Jun 2026
How to know whether your n8n workflow is actually running in production
Most n8n workflows have no alerting, no structured logging, and no health-check layer. They run fine until they silently stop. Here is how to build observability into automation before that happens to you.
Read
8 Jun 2026
What a Make scenario looks like six months after the person who built it left
Automation debt is not a technical problem. It is an institutional memory problem. Here is what happens to a Make scenario when the builder leaves and nobody documents anything.
Read

Notes on jobs that report success and don't do the work.

How to design an n8n workflow that degrades gracefully instead of failing silently

The Make scenario that runs every night and has never once been checked

What silent data corruption actually looks like in a production n8n workflow

n8n error handling that works in testing silently fails in production

How to know whether your n8n workflow is actually running in production

What a Make scenario looks like six months after the person who built it left