fix: retry on stale ETag

[hoonoh]

I kept facing this error from time to time:

arn:aws:lambda:us-east-1:*****:function:*****:91 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:91 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:21 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:26 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:18 is associating to ***** CloudFront Distribution. waiting for deployed status.
Serverless: Recoverable error occurred (Rate exceeded), sleeping for ~5 seconds. Try 1 of 4
arn:aws:lambda:us-east-1:*****:function:*****:58 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:13 is associating to ***** CloudFront Distribution. waiting for deployed status.
Serverless: Recoverable error occurred (Rate exceeded), sleeping for ~5 seconds. Try 1 of 4
 
  Serverless Error ---------------------------------------
 
  The request failed because it didn't meet the preconditions in one or more request-header fields.
 
  Get Support --------------------------------------------
     Docs:          docs.serverless.com
     Bugs:          github.com/serverless/serverless/issues
     Issues:        forum.serverless.com
 
  Your Environment Information ---------------------------
     Operating System:          linux
     Node Version:              12.18.4
     Framework Version:         1.82.0 (local)
     Plugin Version:            3.8.3
     SDK Version:               2.3.1
     Components Version:        2.34.9

Did some googling and found:
https://www.pulumi.com/docs/tutorials/aws/aws-ts-static-website/#preconditionfailed-the-request-failed-because-it-didn-t-meet-the-preconditions

This is caused by CloudFront confirming the ETag of the resource before applying any updates. ETag is essentially a “version”, and AWS is rejecting any requests that are trying to update any version but the “latest”.

So I am guessing this error happens when multiple function updates are requested into single Cloudfront distribution consecutively.

This patch wraps config generation routine (which includes (re)fetching latest CF ETag) and updateDistribution request into an async function and retries if faced with PreconditionFailed error (max 5 retries with 5 seconds delay).

With this patch, the above error is handled as followed:

arn:aws:lambda:us-east-1:*****:function:*****:97 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:97 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:22 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:27 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:19 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:64 is associating to ***** CloudFront Distribution. waiting for deployed status.
arn:aws:lambda:us-east-1:*****:function:*****:14 is associating to ***** CloudFront Distribution. waiting for deployed status.
received precondition failed error, retrying... (5/5)
Done in 141.27s.

[horike37]

That makes sense 👍
Thanks!

[horike37]

🎉 This PR is included in version 1.1.5 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀