Users detail logs of activities during reservations and trigger actions based on those activities

Hi @Daniel. Thanks for the ideas :) the topic of idle sandboxes is tricky, since our customers use sandboxes in a very wide variety of ways and for different use cases. Although improving utilization is an important goal for us, it’s hard to define what tells us that a sandbox is idle (not to mention a feature that terminates users’ sandboxes in the middle!). This often depends on what is in the sandbox (network devices, physical servers, VMs in private or public cloud, public cloud services, storage arrays, Traffic generators, Test software, various applications, databases and more), how many elements are in a sandbox, are elements private or shared, what activity the sandbox is used for and other considerations. In your example you mention events like running a command or opening a remote session, but for many of our users not running commands or opening remote sessions doesn’t say much about the usage (e.g. sandboxes that are started by automated processes, or users that use the sandbox in a different way). Other parameters like CPU usage or traffic running somewhere in the sandbox are also problematic (not always relevant, require sampling throughout the day, how often? What to sample? how to aggregate? some of our customers use sandboxes for minutes others for weeks…)

Another way to tackle this is to limit sandbox duration by defining a shorter default duration and allow extending if needed. If users don’t use a sandbox they won’t ask to extend, and you could immediately improve the utilization.

It’s also possible to use reporting, tracking sandbox consumption and providing users/groups with weekly/monthly information about their consumption, to raise their awareness (even if you don’t actually charge them for usage, it changes things when they see the “bill”).

I’m not sure an OOB feature could work here, however if it's important for you to find a way to terminate sandboxes that aren't well utilized, it may be possible to find a way to extend the system with custom work and implement a specific solution that reflects your definition of idle and how to handle idle sandboxes (which is something that our Customer Success can advise on).

Will be happy to hear community member opinions and suggestions on this!

Hi @Maya.B

I dont understand why this is tricky since we are just asking you to expand the activity logs and use them for something useful. I understand that there are ways to limit the reservations, but in our case, users tend to abuse those privileges, thus this request. As I mentioned before, we do use BI, but we need more than just telling groups about their usage. We are not in the business of raising awareness, we are in the business of cutting costs and saving resources within our company.

Hope it is clear by now, that we are trying to expand the use or your tool within our company but we do need to your help to make it more flexible. Hopefully your product team and roadmap will align to help us achieve our goals. It is good to know that this is a feature that won't be considered though since we now have a reason to look at other tools

Thanks for your time

Hi @Daniel,
It was great talking to you today and learn more about your use case.

What you are basically trying to achieve is an automated process to terminate Sandboxes once an idle time is expired. Each group of users will have a different idle time limit.

Your definition for an idle Sandbox is "no resource or Sandbox commands were executed nor application-links operations (e.g. telnet, ssh, RDP, VNC) were used".

The only gap that blocks you from implementing such a script is that application-links are not being registered in the Activity Log today. If we would add such events to the Activity Log you will be then able to develop a script that will utilize the Activity Logs APIs and calculate the Idle Time for each Sandbox, then terminate any Sandbox that exceeded the predefined maximum for the specific group the user belongs to.
I hope this help to clarify the request. Our product team will review this comment and will provide their feedback. Have a good day!

So there is no hope to get this as an OOB solution and a customized solution is the only way? we are getting asked to figure out the utilization of resources within an testbed for the reasons behind so we need to know ASAP which way, whether is customize, velocity or any other way to accomplish this

thanks

Hi Daniel,

The immediate solution would be to have a script that will utilize our Sandbox APIs to monitor the sandbox for activity and take actions based on rules you will define.

I will be happy to have a meeting with you and your team and guide you with the implementation steps.

Feel free to reach out to me at your convenience to schedule this meeting.

Thanks,

Alon