Hacker News Reader

Comment by pjungwir | Hacker News Reader

I don't see anything here about replication or even backups. To me that is why I'd use hosted Postgres. RDS gives me multi-AZ failover but no access to the slave for scaling read-only queries, no way to configure pushing WAL archives, etc. (They do offer slave access for MySQL, so hopefully this is just a matter of time.)

Going up against RDS, Heroku Postgres, and the dozens of existing hosted Postgres vendors is a tough market. You have to offer more than just `apt-get install` and some tweaks to postgresql.conf!

View on HN

Hello, Database Labs founder here. As others mentioned, we do continuous offsite backups (to Amazon S3) via WAL logging, as well as daily snapshots. I'll be updating the features list soon based on feedback we've gotten here to make that more visible.

We do have replication setups, but it's currently only available via manual operation / support ticket. We'll be adding controls for that to the user interface soon.

Besides the 'apt-get install' and a few config tweaks, we are offering a managed service. You are getting a professional operational staff on call 24 hours a day to ensure that your database server is always running, always has the latest security patches, that the backups always happen on schedule, and so on.

I think the ongoing value of old fashioned system administration is easy to overlook until you wake up one day and find it crashed or you were hacked at 2 AM and they have a pile of angry customers!

It's easy to see the value in a "positive result" proposition -- pay and you get something. It's a little harder to see the value in a "negative results" proposition -- you are paying so that something does NOT happen; for proactive work to ensure that your database doesn't die horribly and unexpectedly. If all goes well, it just works.

We are getting good feedback from people who have either done sysadmin stuff and know firsthand how much work goes into tending servers, or who have neglected to do it and then been burned at an inopportune moment by a server failure. These people are usually quite willing to pay for someone else to do it. The value proposition here is that we will handle all that for you.

Thank you for your reply! Sorry if all I offered was criticism. I almost started your company last year with a buddy, so I've thought about it a lot. (We decided it was too risky that AWS would release Postgres RDS--and they did!)

In my comment, I tried not to compare your service to DIY, which of course may work for me but not most. I'm comparing it to RDS and Heroku Postgres. There are definitely gaps you can fill, but you will have to start offering some more advanced features if you want to compete with such trusted players.

If you want to compete on price, there are already dozens of other companies offering low cost, low quality Postgres hosting. So I think your best shot is offering a high quality product.

Also depending on your goals it may be fine if your product earns you more dollars from consulting than from hosting. :-)

No worries, I think it's good to cover all possibilities rationally, and your feedback helps.

We've gone through the same process, too, and come to a somewhat different conclusion. :) We perceive a gap in the market -- as you mentioned, there are many low-end Postgres hosting places, and a few expensive high end places. All are lacking features that we (personally) want, too, in one way or another.

I think that, at least partially, it comes down to Amazon and Heroku having many other lines of business. Postgres is just not a major priority for them at all. By focusing entirely on Postgres and only Postgres, we can deliver a better experience.

Competing on price is a race to the bottom. I don't want to be wading into a commodity market trying to pick up a million $5 customers. I'd much rather go for the higher end segment where we deliver premium service and features for people who see the value in that and are willing to pay for it.

I'd love to chat with you sometime in more detail about the Postgres idea.. if you're ever in San Francisco, I'll gladly buy you a beer or a coffee!

Congrats on taking the leap and going for it! Best of luck to your company.

No plans to visit the Bay Area right now, but maybe I'll be down for a conference some time. If you're ever in Portland, I'll do the buying. :-)

> Fully backed up continuously, offsite, via WAL logging

Thanks, I missed that!

Replication and backup solve different problems. What if I accidentally drop a table and need last night's backup? Are they archiving the WALs too so you can do point-in-time recovery?

Hi, Database Labs founder here.

We do nightly snapshots plus ongoing continuous WAL logging, so yes, you can do point-in-time recovery to any moment in the retention window. The retention window is 14 days minimum, with longer windows available upon request.

That sounds sensible, thanks; I'd misunderstood and thought you were doing continuous replication to another PG server, now I realise you're archiving WAL to S3.

They do mention offsite backup using WAL shipping.

Reply on news.ycombinator.com