On Tue, Feb 18, 2014 at 4:11 PM, Jonathan Vanasco <[email protected]> wrote:

> I just want to bring this up, because the Pylons/Pyramid error reporting --
> and the SqlAlchemy reporting -- can often make an error look to be in a
> different spot than it really is

It looks like it might be what Jeff was talking about.

> Sidenote: How are you liking the pg RDS support?

I like a lot of AWS products and we use a lot of them. We also use a
lot of them really really well and they work flawlessly with DynamoDB
being the best of the best.

PG RDS is fairly bad, we're going to move off of it because it's
really unusable but we're creating our own replacement for it. But if
you enable Multi-AZ support you'll have a ton of latency that's
unexplained (multiple seconds of latency on each query or connection).
Sometimes write latencies will spike to multiple seconds and so on. We
ran a pg 8.4 database in ec2 on our own instance with EBS and
everything and never had these issues a single time in 5 years. We've
had more problems with PG 9.3.2 in RDS in 4 weeks than in 8.4 on our
hardware in 5 years. Still vastly immature IMHO, we're using PIOPS and
EBS optimized instances.

The other issue is that there are no c3s in RDS yet so you get a much
faster machine (5x) by running your own c3s instead of the m2s for 50%
more cost.

-- 
Valentino Volonghi | Chief Architect
P: +1-877-723-7655 | F: +1-415-738-8026
AdRoll.com Increase Sales. Raise Awareness.

-- 
You received this message because you are subscribed to the Google Groups 
"sqlalchemy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sqlalchemy.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to