Make with Astro

Three things that commonly go wrong.

1. 01

   Failed runs need a recovery and notification plan

   Error handling is a thing you build rather than a thing that exists. Where a scenario has no route for failures and nowhere to announce them, the first sign of one is a gap in the data.

2. 02

   Credit usage grows with the work each scenario performs

   A scenario iterating over a collection does more work per run than the module count suggests, and usage depends on which modules and services are involved. I estimate it across the branches, the loops and anything billed by the work it does, before the scenario is switched on.

3. 03

   Webhook order is not guaranteed

   Two submissions within a second can be processed out of order. Anything where the second write depends on the first needs to be built for that rather than assume it.

What is included.

• Form submissions and content changes into Make scenarios

• Routers, filters and enrichment steps

• CRM, email, Slack, Airtable and database writes

• Error handlers for every important branch

• Duplicate prevention where a retried step could write twice