Fixing hard to reproduce app bugs using automations

The goal of this short topic is to show that while automations can be used to automate (simple) data processing and business process/rules, they can also be used for app debugging.

From my experience, many long-lasting bugs in a Tulip app manifest themselves by an incorrect value inserted in a table at some point. Sometimes, it’s very hard to reproduce them because the conditions to do so are quite specific/obscure, which can be made even worse if the app wasn’t well built.

Here is what I suggest:

  1. Create an automation that reacts to records being added/updated in the table.
  2. Examine the record’s values to find out if something is wrong.
  3. If so, trigger a notification (e.g. send an e-mail) to the person responsible for app maintenance with relevant information if needed (and/or log anything useful in a debugging table).
  4. Quickly contact the person who used the app and ask that person what exactly they were doing.

This technique has the following advantages:

  • You can quickly get information from final users to know how to reproduce the bug.
  • No matter which app triggered the issue, you will be warned.
  • This will potentially reduce the time you spend debugging manually.
  • This can notably help if you are in a situation where rebuilding the app from scratch is too costly (in time or money) and manual debugging has failed so far.

Note that this doesn’t depend on a specific Tulip plan since record-triggered automations are available in all of them.

1 Like

This is a great use case for Automations! Thank you for sharing! We are also working on a way to surface trigger errors in the product, but Automations are still the best way to do data audits and other data sanity checks.