external-dns: Too slow to sync with 400k records in route 53

Hi, there. I am testing external dns with Route53 of AWS provider. I have two environment; production, dev. And then I checked external dns in dev worked fine but didn’t in production. As I see log messages, I found too much time is required to sync every time. It’s totally different value compared to dev environment.

time="2020-09-15T08:06:35Z" level=info msg="All records are already up to date"
time="2020-09-15T08:41:24Z" level=info msg="All records are already up to date"
time="2020-09-15T09:16:30Z" level=info msg="All records are already up to date"
time="2020-09-15T09:51:29Z" level=info msg="All records are already up to date"

I guess that the reason is that the number of records in production is really huge so that the time to cache records highly increases. I am not sure because I didn’t carefully check the code. The number of records in my zone is around 427k. Can this affect the external dns working well?

Please let me know how to fix it.

About this issue

Original URL
State: closed
Created 4 years ago
Comments: 19 (8 by maintainers)

Most upvoted comments

So there is a gap of 30min between:

time="2020-09-18T08:15:31Z" level=debug msg="Considering zone: /hostedzone/REDACTED (domain: test.com.)"

and the next log:

time="2020-09-18T08:54:03Z" level=debug msg="Skipping endpoint foo.test.com 300 IN CNAME  bar.sendbird.com [] because owner id does not match, found: \"\", required: \"<REDACTED>\""

if i understand correctly it spent 30min talking to aws to check the records aws already holds. I must reread the code

Benjile on Sep 18, 2020