Hm, but for this you can disable this setting?
Looks like a bug, we will check
Hello @helloitse
At some point we stopped deleting entities in some cases and instead we just set “Deleted in HubSpot“ field to true. After the merge old entities will have it as “true”. Then you can either delete them manually or setup some automation. Please note that any rich-text annotations or custom fibery data won’t transfer to new records automatically 
In case the field is not updated, please trigger “Full sync”
Out of curiosity, does the Jira two-way sync make use of the normal Custom-Integration API with some extra, unlisted endpoints/commands/parameters or was it a totally custom one-off?
I completely understand the team not prioritizing more integrations like this, but I’ve always wondered if there was some undocumented API stuff that we could experiment with. I have successfully gone down the rabbit-hole of Custom-Integration db → editable mirror db → send changes via external actions → get update via webhook (plus a bunch of little tweaks to make sure there is no looping and everything is linked correctly), but it’s pretty clunky and I would happily spend the time doing some reverse engineering if I knew the was a chance I could build something similar. I am aware that it would require some non-trivial coding to orchestrate and prevent looping.
I would also love to know this! A lot would be possible for us to connect if the flood gates were opened to bi-directional sync like in Jira. @Chr1sG, do you know who the right person to ask would be?
It was a custom implementation - there are no universal endpoints that we’re choosing to hide from users.
I’m afraid there are no flood gates to be opened, sorry.
Are we, as users, able to make a bi directional custom integration some how?
When you say custom implementation, does that mean it doesn’t use the Integrations framework that exists?
Correct. It does not use the standard integation framework.
No. Not possible.
Does it use some part of it with a different integration loop under the hood? I may have done quite a bit of reverse engineering on some of the undocumented features of the native integrations and managed to make the standard integration system give me writable fields like the Jira integration does (please don’t patch 
).
Fibery team, in lieu of a real Bi-Directional API, would you consider adding a few features to the existing Custom Integration API that would make it much easier for users to build their own?
I have spent quite a lot of time working on this problem and I think with a few additions, the current Custom Integration API could actually be a perfect base for custom Bi-Directional sync.
Features/Edits
- Document how the writable fields and dataProviding work in the Jira integration.
- Allow for new entity creation in integration DBs (as an option through the integration config).
- Emit requests for entity Creation and Deletion the same way that the Jira integration does for Updates.
I think this would be the minimum to enable a full featured Bi-Directional sync, but would hopefully be much easier than a full Bi-Directional API. I will happily make a separate feature request, but I wanted to see if Fibery team would be open to these ideas.
We will consider that, I will talk to devs this week and we will see how it can be done