Sharp uptick in "invalid response from upstream server" errors

6 replies
Last updated: Jan 19, 2021
Hi, just reporting a sharp uptick in error messages in a project. Two separate services are reporting
invalid response from upstream server
. More details in 🧵 .
AI Update

I understand you're experiencing "invalid response from upstream server" errors across multiple services. This type of error typically indicates an issue with Sanity's infrastructure or connectivity between services.

Here are the steps I'd recommend to troubleshoot and get help:

Immediate Actions:

  1. Check Sanity's Status Page - Visit https://status.sanity.io to see if there are any ongoing incidents or degraded performance reports. Upstream server errors often correlate with platform-wide issues.

  2. Check Which Services Are Affected - Since you mentioned two separate services are reporting this, identify:

    • Are these errors coming from API requests (apicdn.sanity.io or api.sanity.io)?
    • Are they related to Studio, asset delivery, or Content Lake queries?
    • What specific endpoints or operations are failing?
  3. Review Your Network Path - "Upstream server" errors can sometimes indicate:

    • CDN or proxy issues between your application and Sanity's infrastructure
    • DNS resolution problems
    • Temporary network connectivity issues

Gathering Diagnostic Information:

For your thread, include these details to help diagnose:

  • Exact error messages and HTTP status codes (502, 503, 504?)
  • Timestamps when errors started
  • Your project ID (if comfortable sharing)
  • Whether errors are intermittent or consistent
  • Any recent changes to your configuration or deployment

Getting Help:

Since this appears to be an infrastructure issue affecting multiple services:

  • Post in the Sanity Community Slack #help channel with your details
  • If you're on a paid plan, contact support@sanity.io directly
  • Check the Activity Feed in your project to see if there are any unusual platform activities

The API CDN is globally distributed and designed for high reliability, so widespread "upstream server" errors are unusual and likely indicate either a temporary platform issue or a specific configuration problem that support can help resolve quickly.

Keep us posted in your thread with what you find - this helps others who might encounter similar issues!

Show original thread
6 replies
browser-based service reporting via sentry.io :Rose to x200 reported issues and flattening out at 150 for every ten minutes over the last half hour

UnhandledRejection: {"response":{"body":{"message":"The upstream server is timing out"},"url":"https://.<http://api.sanity.io/v1/data/query/new_production?query=%7B%0A%20%20%20%20%22access%22%3A*%5B_type%20%3D%3D%20%22person%22%0A%20%20%26%26%20email%3D%3D%24userEmail%0|api.sanity.io/v1/data/query/new_production?query=%7B%0A%20%20%20%20%22access%22%3A*%5B_type%20%3D%3D%20%22person%22%0A%20%20%26%26%20email%3D%3D%24userEmail%0>...
EDIT: edited out projectId once I saw the issue was fixing itself
🙂 If you peeps@sanity want it for follow-up, dm me and I’ll send over details.
This seems to be flattening out. I can use the service myself now, even if reports in sentry are still ticking in. I was possibly a bit quick on the trigger here and could have waited longer in the hopes of it resolving itself, but the issue-rate triggered my critical alert in a mission-critical service for a customer. 😄
Thanks Haakon, we're on it! Will keep you updated while we try to resolve this asap 🙂
The flow-rate of issues reported on my end seem to be flattening out now fyi ( ie, curving downwards to normal levels ) 🙂 You’re either very quick on the fix, or you have self-regenerating servers 😄 Thanks for the quick response in either case ❤️
We should be in the clear indeed, although we're still monitoring. Really appreciate the extra context and updates! 🙌

Sanity – Build the way you think, not the way your CMS thinks

Sanity is the developer-first content operating system that gives you complete control. Schema-as-code, GROQ queries, and real-time APIs mean no more workarounds or waiting for deployments. Free to start, scale as you grow.

Was this answer helpful?