Hi Dominic,

Currently we set all TTLs to 1800 which is 30 minutes so you shouldn't be seeing a 2 hour TTL unless the entry was DNS cached somewhere in between.

Your suggest on lowering the value when a new record is created and then having it automatically raised is very interesting, the only issue would be editing an existing record it would still be under the default 1800 seconds and so it would take up 30 minutes for it to expire in various locations.

For DNS testing we usually just update the records and clear cache locally and then query the nameserver directly to make sure they are valid.

We have thought about opening up the TTL for customers to set so that could be interesting. If it gets enough votes and discussions we'll move in the direction the customers want. =]

Hi Dominic,

Currently we set all TTLs to 1800 which is 30 minutes so you shouldn't be seeing a 2 hour TTL unless the entry was DNS cached somewhere in between.

Your suggest on lowering the value when a new record is created and then having it automatically raised is very interesting, the only issue would be editing an existing record it would still be under the default 1800 seconds and so it would take up 30 minutes for it to expire in various locations.

For DNS testing we usually just update the records and clear cache locally and then query the nameserver directly to make sure they are valid.

We have thought about opening up the TTL for customers to set so that could be interesting. If it gets enough votes and discussions we'll move in the direction the customers want. =]

I would use this extensively and it would greatly reduce my admin burden!

Also, the $TTL in the zone file (according to the labs panel) is 7200. You might want to check with engineering @muretsky

Yes, my zone file is set to 7200 as well. How can this be updated to 1800?

I can stand up a replacement server in about 300 seconds, or 60 seconds if I have a fresh image. These TTLs are crazy long. But the worst part is being told the wrong information over and over. DO support doesn't understand DNS and resorts to blaming the customer rather than looking into the issue.

Moisey, the zone file in DO's DNS interface says 7200 seconds, not 1800 seconds. `dig` says 7200 seconds, not 1800 seconds. This is broken at least for some customers and all I'm seeing from DO is how I'm doing it wrong.

We're going to lower the default this week to 600 (10 mins) this week and then look to add custom TTLs later in the future.

I agree with TTL's being far too high and non configurable is tough to work with. I can spin up a new server in 7 minutes so would be great to flick my changes over in a timely manner. Please lower the TTL's or make them configurable as soon as possible.

No other issues everything else is great! :)

I'd like to see the ability to set each record's TTL separately. This and having DNS changable from the API are two of the things currently keeping me from moving some servers from Linode to DO.

This is an important setting, especially when your use to having it. I would like to see access to this in the near future. Especially when moving domains around and dev work.

Hello. It's been many months since the adjustable TTL discussion started and was just curious if there was any progress on this? We're working with a few new clients and before we direct them to DO we wanted to make sure this was an option. If this is not going to be an option, can anyone reccomend a provider you've had a positve experience with that does also offer this option? Thanks

Thanks for the update, Mr. Uretsky. I'm looking forward to having this feature added! Lowering TTL values is critial for certain domains I host as they have several A records for failover. It is a common and extremely useful feature for most. Even if this were added as an API call for now, I as well as others could deal with that.

Probably the single biggest thing stopping me using DO over Linode for a load of clients.

Any chance of an update on where you are with this?

Just voted.

This could be great (missing it after some DNS based idea we were planning).

The Fisher-Price DNS panel and the lack of IPv6 are the main things holding me back from leaving Linode entirely. Looking forward to this being implemented!

Definitely check out how Linode does DNS config, as that amount of control is nice to have.

Wish I could give all 10 votes here. RackSpace has 300 second for the default TTL and I can set it even smaller. 300 seconds is ideal. Current 1800 is better then two hours but still 6 times more then RackSpace. I think 300 is reasonable. I can imagine applications that will require smaller value.

will be great that function! with the api

I hope that in the future one could choose TTL by default or what you want in control banell on the Digital Ocean

Just voted. You guys should at least lower the default value to 5min, like Rackspace. Is crucial for me that the information about new CNAME records propagate as fast as possible. So I'll probably have to wait a bit more before moving from Rackspace.

+1