SERVFAIL resolving

I’m frequently getting SERVFAIL resolving Amazon S3 hostnames, seems to be an issue on cloudflare’s side.

Connection info:

My test results:

$ ./ @ - 0/1000 failed @ - 49/1000 failed @ - 0/1000 failed

And my test script:

dns_test() {
    failed=0; try=1; dns=$1; hostname=$2
    for try in $(seq 1 1000); do [ -z "$(dig +short +tries=1 $hostname @$dns)" ] && failed=$((failed+1)); done
    echo $hostname @ $dns - $failed/$try failed

1 Like

seeing this behavior as well, seems to have started yesterday at approx 17:20p PDT

only certain subdomains of seem to be affected

1 Like

Sorry about that, I’ll take a look.

1 Like

Currently working for me from EWR using DNS over HTTPS. @noc9 Do you know which edge you’re hitting?

the servers in question that are affected by this are in us-east-1 and hitting IAD

1 Like

We are also seeing this issue when trying to resolve us-east-1 S3 bucket URLs (e.g.

It seems like several bursts of traffic were ratelimited too aggressively which formed a feedback loop with increased retries in some places. It should be cleared up in IAD now.

thanks for looking into and fixing this!

Glad I found this thread. Thought I was losing my mind all day, checking DHCP server, switch, ISP, etc. Thanks for resolving the issue!

Is this happening again? Can’t connect to out of us-east-1.

Contact Amazon or… Have problems with *Read Me First*?