JavaScript is off. Many parts of this site require JavaScript to function properly.

How Unlimited is Unlimited?

Oct 16, 2017 6:37:00 AM / by Jeff Knepper

The second edition of a weekly Question and Answer column with Gary Stern, President and Founder of Canary Labs. Have a question you would like me to answer? Email

Dear Max,

That's a great question.... without a simple answer.  We reference this package as unlimited, but as we all know, there is always a limit.... though we doubt you will ever find ours.  Let me explain.

A Canary Historian is designed to gather time-series data for many different industries with hundreds of applications.  Every user has different data with different parameters, and the Historian is optimized to work in each of these situations.  For instance, one user may be collecting only 50,000 tags at one second scan intervals, with each tag value changing on average once ever other scan.  Another individual may be collecting 500,000 tags at ten second scan intervals with an average tag change only once ever five scans.  Now factor in the performance of the hardware that is housing our platform.  As you can see, there are simply too many variables, most of which are constantly changing.  As I tell the staff all the time, "real data is messy".

So how do we claim an unlimited server when we have no way to know exactly how you will ask it to perform?  We simply over engineer everything.  Internally, we have tested our database in two different "overload" scenarios.  In the first, we logged a million tags and changed all million tag values every second.  The database ran without issue.  Secondly, we tested a single machine with 25 million tags, with a much higher scan time and a much lower change rate.  Again, the database performed.  Obviously both of these tests were conducted using higher end hardware, I used my top of the line desktop machine for one of them.  You can read more about other aspects of our performance testing here if you like.

Here is the takeaway... we have purposely developed a database that should be able to handle millions of tags on a single server and provided a licensing model that will allow you to add tags and allow anyone in your organization access to the data without worrying about future costs.  For extreme use cases we will gladly make recommendations on both server capacity as well as hardware recommendations, just let us know your specific situation.


Gary Stern
President and Founder

Have a question you would like me to answer?  Email


Topics: Ask Gary

Jeff Knepper

Written by Jeff Knepper