Initial shop sync failed - what to do
Hi, when a shop installs, I run an initial sync of products, variants, collections, productMedia... Sometimes this sync fails.
E.g. just now failed with 28 errors, all of them being
GGT_INTERNAL_ERROR: Unexpected HTTP error from sandbox: Response code 502 (Bad Gateway)
TraceID: 3878f6febfa876b1a21fc04bcc3c07a3
What's the recommended course of action?
- The sync ran 50 minutes (pretty big store). Triggering a complete re-sync would be quite expensive.
- Can I assume that data was correctly synced even though we had a few errors along the way?10 Replies
We'll take a look.
Just wanted to update you. We know what's happened, the VM that two of your sandboxes were running on crashed. Because of this there was a brief window where we continued to send requests to these unavailable sandboxes. Our system is designed to see this unavailability and send requests to the remaining available pods. I'm still digging into why there was a ~5s window where we did not route to available sandboxes.
The error was transient though if you want to retry the sync
Got it, thank you.
@[Gadget] yandu Just want to let you know that I'm hitting similar issues with my background actions as well. Seeing lots of 502 bad gateways (GGT_INTERNAL_ERROR). Appreciate you looking into this 🙏
@Barkos do you have a traceid and your app's name?
@[Gadget] yandu
Most recent trade id: 23b9d3c74177866ef33deb2396646248
app name is brand-scout
Side note, please note this message: https://discord.com/channels/836317518595096598/958764380038782976/1439985658218283109
@Barkos I'll send you a DM
@[Gadget] yandu For anyone watching this thread - Yandu successfully helped me identify my problem and get it resolved 👏 . It was likely an OOM issue, I was loading to many records in memory.
We hope to deploy a fix for the other 502 issue that @Simon saw tomorrow
@Simon the fix finally shipped. Let me know if you see any more failures due to 502s and we can investigate.
Will do, thank you @[Gadget] yandu