Reltio Connect

Hi all,

We have some questions related to your day to day operations

To give you some context : 

We plan on scheduling on a daily basis dataloads (full updates) from 12 differents files and the total amount of loaded records will be around 7.5M.
Those records are then matched by our matching rules, we also have a job that updates records based on the matches found so if the records has a potential match that attribute will be changes to 'Yes' and if not it changes to 'No'. 
The records are then sent to Salesforce via the SF Connector based on some filters.

Problems that we face : 

The dataload of all our files takes quite some time to fully load. 
Everytime we dataload the files, millions of events are being sent to the SFDC queue so the queue needs to process all those events before it can send the records to Salesforce. From what we saw when we monitored the amount of events in the queue, the processing time is around 125K events per hour so the updates and creation are not sent to Salesforce the day they are received in Reltio because of the processing time

For example :

I am dataloading on Monday the 7.5M records, new records have been added to the file and the dataload generated let's say 5M events to the SFDC queue. 
In order to process the 5M events based on a velocity of 125K events per hour, it would take 40 hours to process all the events meaning that it would take 40 hours for the new records to be sent to Salesforce.

Questions that we have : 
- How much time does it take for you to load your data in Reltio ? Our files sizes vary and can have around 20 attributes for some and 150 attributes for others, but how long does it take for let's say 1M records ?
- If you use the SF Connector, do you know how many events can it process ? In 24 hours ? (even if you have an estimate, that would be helpful!)
- What are the processes that you run on a daily basis / weekly basis so you do not impact the performance of your tenant / of the queue ? We are wondering if the processing time is slower on our end because of how we configured our tenant

Here is a link to a diagram of what was described above : https://excalidraw.com/#json=oh1zXLIUgH0jgtdD5Zakh,ulRwR1ZMRd3ER9lX0RW5Lw

Thank you all for your time,
Léon Tang

Hi Leon, apologies for the delay in responding.  Unfortunately we don't have any performance metrics to share with you just yet.  We are looking to get some benchmarks published.  As soon as we have some data available, we'll be sure to share it.