Url returning 404 for batch uploads

[ChrisEllis007]

• analytics-react-native version: 2.20.4
• Integrations versions (if used): @segment/sovran-react-native: 1.1.3
• React Native version: 0.76.7
• iOS or Android or both? both

Steps to reproduce

Create a config:

const segmentClient = createClient({
writeKey: SEGMENT_ANALYTICS_KEY,
trackAppLifecycleEvents: true,
});

When sending a batch the url visited returns a 404:

Expected behavior
Url should be valid

Actual behavior
Segment returns a 404

Workaround
I have found a workaround is to use a proxy and disable segment endpoints. This fixes the batch issue, but it may introduce others.

const segmentClient = createClient({
writeKey: envConfig.env.SEGMENT_WRITE_KEY,
trackAppLifecycleEvents: true,
useSegmentEndpoints: false,
proxy: 'https://api.segment.io/v1/batch',
});

[leethree]

we are seeing the same issue in latest version. is it caused by #1043 ?

[wenxi-zeng]

sorry all. this issue is caused by #1043. @sunitaprajapati89 is working on the fix. in the meanwhile, please pass https://api.segment.io/v1/b as proxy in config like @ChrisEllis007 suggested or rollback to the previous version. will ship a fix asap

[sddanila]

I am seeing a similar issue and it resulted in substantial data loss.

For the Segment team, the issue is coming from analytics-react-native/packages/core/src/plugins/SegmentDestination.ts:

• The issue was that we did not have a proxy defined (we were using the default url).
• when uploading events (line 59), the url is populated by this.getEndpoint, which is returning https://api.segment.io/v1/b
• this function gets the endpoint from the analytics config which always appends the /b where no proxy is defined
• the segmentClient was trying to post to https://api.segment.io/v1/b/b which was returning a 404
• this started happening without us having made any changes to our Segment setup
• the url that is getting returned upon creating the Segment client changed on March 4/5, 2025 and is returning a value that does not match the expected url (https://api.segment.io/v1)

I resolved this by overwriting the default with a proxy in the segmentClient to proxy: 'https://api.segment.io/v1/b'

[leethree]

As this bug is already in our released production app, we are already losing data for users on the 2.20.4 SDK. Is it possible for Segment to update their servers to accept https://api.segment.io/v1/b/b?

[wenxi-zeng]

hi @leethree, so sorry about that. can you check if you have regional settings in your source settings? if so, please follow the instructions below and see if that resolves the production issue. according to this line, the issue should be solved if the remote settings provides a value for apiHost. this is also the reason the issue slipped away from our manual tests, because newly created sources now have this value but the older ones have it as null

1. go to your source on segment dashboard -> settings -> regional settings -> Choose Data Ingestion Region
2. select your region (Oregon if in U.S.)
3. save and test to see if the issue resolved

[leethree]

@wenxi-zeng It dons't look like we have that setting in our dashboard as we are on the "Team" plan, sadly.

[wenxi-zeng]

hi @leethree and all the customers who got impacted by this but do not have access to regional settings, please send me an email (removed) with your writekey, source ID and region (US or EU), our team will patch an apihost value to your source settings, so your production issue could be resolved. thanks!

[wenxi-zeng]

hi all, the issue should be fixed in 2.20.5. please have a try and let us know if you still experience issues.