# Custom Integration drops all fields

**URL:** <https://community.fibery.io/t/custom-integration-drops-all-fields/11427>\
**Category:** Get Help\
**Created:** [October 15, 2026, 4:02pm UTC](https://community.fibery.io/t/custom-integration-drops-all-fields/11427 "2026-10-15T16:02:11Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![Tommy\_Hedley](https://sea2.discourse-cdn.com/flex020/user_avatar/community.fibery.io/tommy_hedley/32/6468_2.png) [@Tommy\_Hedley](https://community.fibery.io/u/Tommy_Hedley)\
**Post date:** [October 15, 2026, 4:02pm UTC](https://community.fibery.io/t/custom-integration-drops-all-fields/11427/1 "2026-10-15T16:02:12Z")

</div>

What is the specific mechanism that makes Custom Integrations reset all of their fields?

I’ve been working on an integration with lots of fields and as I refactor, the integration periodically chooses to set every optional field to ‘Skip’, even when I’m not making schema changes. This takes a ton of time to fix unless I just delete the integration and reinstall, but I know this isn’t great either as I can eventually hit the schema limit and no longer be able to make fields.

I’d love to know how to avoid this problem.

---

<div class="post-metadata">

**Author:** ![mdubakov](https://sea2.discourse-cdn.com/flex020/user_avatar/community.fibery.io/mdubakov/32/10_2.png) [@mdubakov](https://community.fibery.io/u/mdubakov)\
**Post date:** [October 15, 2026, 4:39pm UTC](https://community.fibery.io/t/custom-integration-drops-all-fields/11427/2 "2026-10-15T16:39:06Z")

</div>

Looks like a bug, we will check asap

---

<div class="post-metadata">

**Author:** ![Oleg](https://sea2.discourse-cdn.com/flex020/user_avatar/community.fibery.io/oleg/32/1887_2.png) [@Oleg](https://community.fibery.io/u/Oleg)\
**Post date:** [October 15, 2026, 4:48pm UTC](https://community.fibery.io/t/custom-integration-drops-all-fields/11427/3 "2026-10-15T16:48:35Z")

</div>

Hello, @Tommy_Hedley

Could you please double check that you didn’t change ids of fields or type ids in synchronisation schema? The described behaviour happens when the schema/fields are changed.

Thanks.

---

<div class="post-metadata">

**Author:** ![Tommy\_Hedley](https://sea2.discourse-cdn.com/flex020/user_avatar/community.fibery.io/tommy_hedley/32/6468_2.png) [@Tommy\_Hedley](https://community.fibery.io/u/Tommy_Hedley)\
**Post date:** [October 15, 2026, 10:56pm UTC](https://community.fibery.io/t/custom-integration-drops-all-fields/11427/4 "2026-10-15T22:56:58Z")

</div>

Ah, a bit embarrassing on my part 😅 . I was doing some AI refactoring and it changed the case on a single letter on a field replicated on all dbs. I totally missed it on my first pass. It did still set every field for every db to skip, which seems a bit unnecessary, but it was due to a schema change.