How to group individual windows machines into a single 'customer'

I am integrating the Windows SDK with a suite of traditional desktop applications. We have deployments scenarios for this on-premise software that are not trivial. A given 'installation' will reside in an office and will consist of 1+ desktop machines (as well as 0-1 local servers). We consider such an installation a single customer and will want to compare customers against each other whilst still gathering data from each individual workstation. I am currently conducting a trial to see how well Countly can integrate with this model. I have a few questions I am hoping to gather opinions and best practice on below please…

  1. Grouping all ‘workstations’ together in an installation to represent a single customer. What is the recommended practice for individually identifying these Windows machines within Count.ly but at the same time being able to group them together as a single 'user/customer' whilst being able to compare customers?

  2. The suite consists of several applications. What is the best way of indicating which application in the suite is running the current session? I have considered using a single app key for the entire suite, or one per application but am unsure of the implications either way. For the former I have played with encoding the application name within the event name.

  3. Dashboards - I may want to compare events across multiple applications. Is the restriction when adding a widget to only come from 3 apps hard-coded or configurable? what if I wanted for example, 10 applications and be able to chart their combined performance is that possible?

0

Comments

1 comment
  • Hello,
    Let me try answer your questions

    1. Grouping machines into single customer can only be achieved with SDKs using same device ID. So if you have your customer, then you probably have customer ID or license ID, and if all workstaitions can get access to that ID, then they can provide it to the SDK and have all data and all events as single Countly user

    2. If you use multiple apps as one, then there will be no way of separating sessions between them, so my suggestions would be to use separate app_key for each application. You can distinguish events by prefix, as you said, but with sessions is not entirely possible. But if you really want to with single Countly app for all software, then probably your best bet would be to sacrifice some metric, like device, or platform, since you already probably control that for customer. There when specifying session start, you would provide application name as one of the metrics, and in the metric breakdown, you should be able to see session breakdown for each app. But it would also require modifying SDK too, to override the metric. And it has not been tested and is just theoretical, but I think should work

    3. So you can compare apps (up to 10 in app compare section) but it can only compare sessions, users, etc. Not events.
      You can compare events (also up to 10) under Events section, but only within the same app, so it would work if you have single app for all software.
      As far as I know there are no technical limitations to comparing events across different apps, but rather, requires much more work on UI, design and user experience, to decide where to put it in a single app oriented dashboard etc. And none of the clients really requested such feature before. So at the very least, it would be possible to do it with custom Countly plugin.
      About custom dashboard widgets, then I think limitation of apps there are because of not making graph too crowded, so also mostly UI limitation.

    0
    Comment actions Permalink

Please sign in to leave a comment.