Anyone else experiencing an issue with the Bubble API Connector just “deciding” to not properly encode URL’s for a few days, then working again? We’ve been using the same unchanged code for our Bubble app for over two months and use Bubble API Connector to get data via REST from a third-party provider.
It had been working great until last week Monday, when it just decided to start adding %25 in front of the first set of encoded characters. For example, what used to be query=mycompanyname%2C%20inc (correct encoding for “MyCompanyName, Inc”) is now coming across as query=mycompanyname%252C%2520inc. It’s as if Bubble is deciding to double-encode the percent sign. To be very clear…we have made no code changes in 2 months. Also, the behavior started last week Monday, then fixed itself near the end of the day Tuesday, and then reared its ugly head again starting yesterday afternoon and continues today.
We confirmed with the third-party vendor, and with a test endpoint, that it is the URL coming out of Bubble that is bad. We also tested multiple prior versions of our app which worked up to 6 months ago, and they all show the same behavior.
Not sure in what context you mean “set things.” We use the Bubble API Connector with
Authentication: None/Selfhandled
Type: GET
Data Type: JSON
URL Parameters: various per vendor specs, none are Private.
Headers: Content key of application/x-www-form-urlencoded, App ID & API Key, all Private
In the creation of the URL, we use the API Connector element, with path(query) set as Arbitrary Text (which is Get name from page URL): formatted as URL, replace spaces by + is checked. Other items per third-party vendor specs.
As I mentioned before, this has all been working flawlessly with no changes to code or platform version for over 2 months. The fails just started happening last week, went away for a while, then came back. When we ran tests against earlier versions of code going back to over 6 months ago, it also showed the same bad-encoding-by-Bubble behavior (even though they used to work).
Settings request as parameters VS directly in URL could help you solve this issue. This is more something to test as a workaround. Because it was working fine before, I agree there’s something on Bubble side that explain this issue. So you need to contact support. S
Yeah, we tested it by repointing the API to a test endpoint (WireMock) instead of the third-party vendor. That’s how we confirmed it was the Bubble output that was bad (vendor eventually sent us before and after input gateway logs from their server that showed the same thing…was working great, and then it just…wasn’t). Have already contacted support, but thought I would ask this forum, just in case anyone else’s app was getting bit by what looks to be a platform bug. Particularly because it took us forever to figure out it was Bubble. We assumed it couldn’t be because we had made no changes in several months.
Except on dedicated instance, Bubble push multiple update everyday. None of them should be “breaking” update (when they are, this should be published as a new bubble version…), but this happen too often sadly. So Bubble can be the problem even if you didn’t update your app…
Support confirmed the issue is on the Bubble side. To wit:
“I appreciate your patience. I apologize for the inconvenience this may have caused you. I’ve been able to confirm that other users are experiencing this behavior, and our engineering team is currently investigating the issue. Our team will be in touch directly once it is resolved. In the meantime, please let us know if you have any additional questions, we’re here to help! ”
”