Date: Wed, 6 Jan 2021 13:15:25 -0600
From: Kenneth Marshall <ktm@rice.edu>
To: Don Seiler <don@seiler.us>
Cc: Joshua Drake <jd@commandprompt.com>,
	pgsql-performance@lists.postgresql.org
Subject: Re: High COMMIT times
Message-ID: <20210106191525.GH7881@aart.rice.edu>
References: 
 <CAHJZqBCsS5wiUemiArXMXR4Oo63_9bBfssrXWieE7X55dYEvNw@mail.gmail.com>
 <CAJvJg-SUfe+Fv0JmvBWgyaPx=J9dzEk9Bg=4JheuVwAdNMj6HA@mail.gmail.com>
 <CAHJZqBBnZ1KOjBtOWA0VPw05EuxrHg5eeBCvO-a8Hm5YPTisyQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <CAHJZqBBnZ1KOjBtOWA0VPw05EuxrHg5eeBCvO-a8Hm5YPTisyQ@mail.gmail.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Precedence: bulk

On Wed, Jan 06, 2021 at 12:06:27PM -0600, Don Seiler wrote:
> On Wed, Jan 6, 2021 at 10:51 AM Joshua Drake <jd@commandprompt.com> wrote:
> 
> Looking at the Azure portal metric, we are nowhere close to the advertised
> maximum IOPS or MB/s throughput (under half of the maximum IOPS and under a
> quarter of the MB/s maximum). So there must be some other bottleneck in
> play. The IOPS limit on this VM size is even higher so that shouldn't be it.
> 

Hi Don,

I may just be re-stating common knowledge, but the available IOPS would
be constrained by how tightly coupled the storage is to the CPU. Even a
small increase can limit the maximum IOPS unless you can issue multiple
relatively independent queries at one. I know no details of how Azure
implements their storage tiers.

Regards,
Ken