Enterprise Solution EGDB

[GoranGobac]

I would like to suggest as an Idea the possibility to analyse referenced Feature Services without restrictions (100.000).
For us Enterprise Administrators it is like a comedy show because sometimes Enterprise Solution looks like Home Solution.
Utilisation of Database is for every Enterprise Customer essential when you are managing Dataflows specially when there is need for multiple Jobs, Triggers and Data Manipulation.
That is why I find always this message that I should use Hosted Feature Service really funny because they are just a weight in such Environment. But great for End users.
Hosted Feature Services don't have Instance Handling/ Heap space Handling / SOI.......
I would really wish that the Enterprise Product owner was more in touch with real use cases from customer perspective. This would greatly improve the practical implementation of this product also in other commercial branches. 
Looking forward to feedback and opinions from other Enterprise Admins

[Brian_Wilson]

AH, so they force you to use a hosted data set, which is basically an unmanaged database that you are not supposed to do anything to like optimize it for your environment, but the database YOU manage that can be scaled and tuned is not usable because of an artificial limit of 100000 records.

I am tired of all the artificial roadblocks. I want to solve REAL problems not work around artificial barriers.

[Scott_Tansley]

Personally, I challenge my clients to ask if analytics via Enterprise is the best choice in this situation. 

Typically, the ArcGIS Server hardware is being used by many end-users, and is also very often a similar size (CPU/RAM) as an ArcGIS Pro workstation.  The ArcGIS Pro workstation is being used by a single person.

The impact of requesting and transmitting 100,000+ rows of data via a referenced ArcGIS Server is going to place a significant load on the server, for a period of time, and effectively become a bottleneck for other users.  It's likely to cause timeouts and unreliability or at the very least poor performance.

The load on the Enterprise Geodatabase is likely to be the same for an ArcGIS Pro session or an ArcGIS Server session.  So, for bigger analytics like this I'd recommend the user have access to a Pro license and do it singularly and without impact on other ArcGIS Server/Enterprise users.

I remember doing some consulting with a BI specialist a few years ago and I complained that the max number of features was limited to 1,000 per page - that's now 2,000.  Their software was limited to 512 results per request.  

[GoranGobac]

Hi @Scott_Tansley I find your insight and posts always helpful but in this case I am not 100% with you on this topic.
First of all if Esri gives this possibility to Hosted Feature Services it should also be given to Referenced Services.
Secondly we had no Bottle Neck with Analysen because we have 3 Server in productive Env. which have enough Power to push through. 
And use +250 Services with +1 GB Heap space
We have daily 100+ GB Data Change. 
Our Spatial Analysis Tool pushes through with 3 Instances.
And Cost Benefit with ArcGIS Pro is not there because the user need like maybe 5% of the ArcGIS Pro functionality.
But as you say I have users where I also say, you need to get ArcGIS Pro.

[Scott_Tansley]

I’m glad you can work with such scale.  Many organisations I encounter are running on resources that are not much above minimal viable product, often due to budget constraints.  Whilst their budgets are small, their ambitions are huge.  Removing the cap would be a source of endless unreliability for them.  

I understand your frustration at the limitation.  But the change would be for all clients and many wouldn’t understand the bottlenecks that result from limited infrastructure and licensing. 

[Brian_Wilson]

I always thought Esri geared the server products for the "enterprise class" which to me would be the biggies like the military or federal government or Forbes 500 and that the rest of us suffered for it.

I also have always wondered how "enterprise" level users could make such fragile software work and assumed it was because they paid Esri to place support staff onsite.

I have the impression of a big gap, me, and (in the US) approximately 3000 other county governments + many thousands more city level underserved because we don't have the resources.

I am thinking this is still mostly true BUT there is another group like the one @GoranGobac is in, at the upper edge of the same big gap, still not enough resources to put Esri staff onsite but hitting limits at the other end.

At this point we do the work with Pro and the other GIS users in our offices rely on web things we do for them and on interacting with us for requests. I can't seriously imagine today telling a planner or an appraiser to learn Pro. We can walk a few public works staff through boiler plate operations to enter data but they have no idea what they are doing. I'm working on a web app for them. Someday.

The gap is also an opportunity for myriad consultants who can fill in one way or another.

I suppose most of Scott's customers are somewhere in the middle between these two groups.