Up now on the new backend! Give it a try!
Discussion
On the old zap.stream, whenever I try to add a server it just buffers interminably. Is that issue fixed?
Add a server? what do you mean by that?
Media servers
Currently we only support NIP-96 on zap.stream for uploads so you can only add a compatible server, it also must return the correct tags during upload or we wont know the resolution
Any servers you recommend besides these two?
Check https://blossomservers.com/ i think anybody running route96 should work
Hi. Users are reporting we can't connect to the API any more. Has anything changed with the connection?
Perhaps URL is different or auth is different?
There is a redirect so it should point to the new api
Shosho app currently can't connect.
Does NIP-42 auth still work with a redirect?
Has the API itself changed? Different endpoints or queries or responses?
Trying to diagnose.
* meant NIP-98 not NIP-42
Yea i dont think i change fix it on my side, its because NIP-98 has the full URL and the host part doesnt match, if you redirect then your NIP-98 event will be rejected, you need to manually update the app to use the new API endpoint
I figured. Is there any possibility to continue to run the old endpoint until I can migrate?
I am AFK for a week.
I am attempting to connect to the new API
"zap.stream/api/v1/account" With NIP-98 auth returns 200 with html body rather than expected API response
I also tried
"api.zap.stream/api/v1/account" returns 404
"api.zap.stream/api/nostr" returns 500
Do I have the right URL and auth?
Is this routing to the API correctly?
The new URL is api-core.zap.stream, there is way to use redirects because nip98 has the full URL in the auth event
No, the data was migrated to the new database, the old one is readonly
I will have to disable host name validation by modifying the backend maybe, checking
I have tested and can connect with the new url.
If the old API is no op, and the new API has the same endpoints, then serving the new API from the old domain would solve my issue.
Is that possible?
I think its up to your client code to follow redirects with Auth header, i think it has to be done manually, the reason for the redirects is that the api path is different on core, ill try to do something on my side first, if that doesnt work for you you can use the new api URL
It’s amazing!!! 🙏💜
Two more issues
1. Lightning top-up does not seem to add to balance. Tried with two 100 sat invoices, invoices were generated and paid but balance stayed at zero.
I see notes that you have fixed this for ZS users but as far as I can tell the error still applies to the invoices created via the API.
2. Streams fail to connect. Shosho will initially connect, ZS will publish the go live event, and then the stream will immediately disconnect.
Struggling to diagnose this error.
ZS uses the api, streams need to connect twice in some cases because of the ddos protection network. Immediatly dissconnecting means balance is negative, i see a lot of that in the logs
Thanks Kieran, two further notes:
1. Shosho is set up to expect that it can stream to free tier with zero balance. Previously you made a specific change to allow this. Is this change still in place?
2. Immediate disconnect also occurs with user nostr:nprofile1qqsfhzff7rw7ljtvn6msmlchamp8sf380t8w0djndl8cgw6e97khj0qpr4mhxue69uhkummnw3ezucnfw33k76twv4ezuum0vd5kzmp0qy88wumn8ghj7mn0wvhxcmmv9uq3qamnwvaz7tmwdaehgu3wd4hk6tcx9qg7eon either tier despite 1200 sats balance.
Created new guest user for testing
nostr:nprofile1qqswqzgl3qcanqxeshl5w8qtn475muw6jq80ntvtmjzq2adr43l8v4czqlllv
Observed:
- connect to API successfully, balance 0
- select free tier
- accept tos
- attempt to connect failed (I guess DDoS protection)
- second attempt to connect temporarily connected and then immediately disconnected (I guess negative balance)
- top up 1000 sats, invoice provided, paid
- top up has no effect, balance stays at 0
VS Expected:
- first connection works
- second connection works
- balance goes to 1000
Invoice: lnbc10u1p5tcfq9pp5eh2rgf544y9edgchrahuxsmyweh6z6kltr2dq0eaemc37e4h3luqdys0fshqtnnw3ex2ctdyp6x7ur4wqsxvmmjyp6hxetjypjnqvpex9nrswpnx9jrjwpsvsunsdtxvc6rwvtrxp3rjephvs6xge33v3snjvpsv4nrjcty8p3xgcecxscr2de4vyekzcehv5mnvdfhcqzzsxqyz5vqsp5u6h4j9e87x9pdqxq6yawfmht9e35lhaaplsjt2lchkwnw0yf6fdq9qxpqysgqyzh42a6h8flcajwkajkazkc7x6ahsh8ke7hemvensf7wdglwhf9x7knkg9lgaxzz0znx8833dsjht4nx4jhqdphdq2vcewyue70wssqqjctpy9
Yea looks like the payment handler is failing at some point and stops running, fixing now in a sec
Should be fixed now
Thanks Kieran
New test result
1. Top up balance now works, thanks
2. Stream with zero balance to free tier still fails. It connects and immediately disconnects vs expected that free tier should work with zero balance -
Can you please reenable streaming to free tier with zero balance?
3. First stream connection always fails regardless of tier or balance. Very disconcerting for users -
Is there a pattern you recommend following e.g. build in instant auto retry?
It should be allowed already maybe there is another check I missed
Changed the backend to allow the old hostname so it should work again for your users, sorry about the issues
Thanks for this.
Tested with new user, old subdomain & path
1. API connection seems to work! 🙏🙏
2. It still immediately disconnects if user has zero balance
I appreciate the effort you're putting in.
In the wash up lmk if you would like me to make any changes or migrations to better fall into the main flow.
Hey nostr:nprofile1qqsx8lnrrrw9skpulctgzruxm5y7rzlaw64tcf9qpqww9pt0xvzsfmgpzfmhxue69uhksmmvv5h8vvrv9e5k7tcprfmhxue69uhhq7tjv9kkjepwve5kzar2v9nzucm0d5hs7f7te7 there’s some regional problem with the streams, we have to use US or Europe VPN otherwise there is no connection, just black. On all clients and all devices ☠️