I don't actually know whether Heroku can failover across an AZ failure.
By default our followers and HA is automatically cross AZ. You also have an ability to create followers across region, but we do not automatically failover on those due to latency.
Heroku's solution is 2 to 4 times more expensive for the same type of DB, and RDS even allows for reserved instances to further lower the bill.
One area is we're focused more on delivering more guidance and expertise around what you're doing with your database, in addition to ensuring your database is healthy and running. An example of this is notifications that we deliver around unused indexes, where you may benefit from other indexes, or other places where you can quickly optimize your DB. This starts to free up a DBA to add higher value tasks or for smaller shops lets you get by longer without a need for a DBA.
Another big area is features we deliver on top of Postgres. This ranges from followers which all you to easily scale read traffic, or allow for your database to be replicated across not just AZs but also regions. There's the other spectrum of this as well including dataclips. Dataclips make it easy to share data in a simple way, as well as build richer dashboards by integrating with google docs, or quickly prototyping APIs.
If you're curious on various technical details we'll be documenting that soon but would be happy to correspond via email, craig at heroku.com
I understand that "forks" and "follows" are easy concepts, stats are cools, etc. but I personally wouldn't want to pay double or triple for that (and I'm not a DBA, so I feel the pain). Not that my word counts much as I'm on the smallest Heroku production plan, so I guess my bill isn't exactly interesting for this kind of discussion. In my case, I would say that even saving $25 off $50 can be offset by these additional niceties. But, I don't know what I would think if I were a customer paying $1000/mo.
Did you add disk? People seldom do, I can attest it's a big part of the bill...
> I don't actually know whether Heroku can failover across an AZ failure.
"Followers" have preferred another AZ for quite some time -- almost since the beginning -- and Heroku Postgres HA is based on similar technology. So, yes.
So pricing wise both double if you go multi-AZ.
I didn't claim to have done a detailed price comparison but a quick look at the small sizes which are currently the ones relevant to me.
No doubt about that. To date, Heroku's product model has preference for fewer, but common choices rather than more flexibility. Heroku's staff sees fit to give in sometimes, but coarsely speaking, that's pretty much how it goes.
Last I heard Heroku is hosted on top of AWS. Does anyone know if Heroku's cheapest Postgres plan is hosted on a dedicated AWS Micro instance, or do they buy large AWS instances and host multiple databases on each box thereby potentially providing more IO performance?