GitHub releases frequently failing

MongoDB Atlas App Services & Realm Atlas Triggers & Functions
#1

Hi

I develop code using VS Code and linked to GitHub which is synchronised with my Realm account. I have on occasion noticed errors during the deployment process (in the Deployment section of the Realm UI), however this has been getting more frequent. Over the past 2 days I have had 4 failures in 10 attempted deployments. The error messages vary, with error invalidating cloudfront being one of them. Attempting to redeploy sometimes works and sometimes doesn’t. Below is a screenshot showing some of the errors. I’m pretty sure it is not an issue with the code as sometimes I just add a space to the code and resync to get it to deploy again.

Is this something I can fix with settings from the UI or code or is it a bug? Thanks

image
image1527×775 140 KB

#2

Hi Simon,

Does that only happen when making changes to hosting files or with other types of changes as well?

Regards

#3

Hi @Mansoor_Omar thanks for taking a look at this. I can’t say with 100% certainty but spot checking a few deployments all of the of the ones that failed first time included hosting changes, whereas many of the ones that succeeded first time didn’t, so inclusion of hosting files might be a contributing factor.

I know from other work that cloudfront can be very slow and timeout sometimes and we’ve needed to put retry strategies in.

#4

if it’s useful here’s the exact error message from a couple more from 24-May.

Failed: error invalidating cloudfront cache: ServiceUnavailable: CloudFront encountered an internal error. Please try again. status code: 503, request id: aa2e6a1d-7941-4d6d-93c6-0c13f4e71ecf

Failed: error invalidating cloudfront cache: ServiceUnavailable: CloudFront encountered an internal error. Please try again. status code: 503, request id: 36d5feb0-92e4-49db-919a-96feb9859687