damies13
commented
9 months ago Hi Dirk,
Thanks for the enhancement request, I remember our conversation about this, but I may not in future, so these are my notes to remind myself as much as anything, correct me if I misunderstood.
anyone else Please add comments if you wish
- You want a way to seperate the monitoring robots from the robots that test the AUT
- This could be by splitting the plan tab into a Plan-Test and Plan-Monitoring tabs (I worry this might be more confusing though)
- This could be by having a Test and Monitoring sub-tabs on the Plan tab
- What about adding a second test row frame (just wondering which will be neater? maybe I'll have to try both ways and see what the usability is like)
Should:
- monitoring robots show on the plan graph?
- be included in the total robots line on the graph?
I'm thinking no to both, currently they do. Your last line also indicates it should be no.
BTW there is already a way to distinguish between agents that run the AUT test cases and agents that run monitoring, you use agent filters for that, and you can use custom properties on the agent to help with this
As I understand it, I use scripts to collect metrics from the systems involved. The scripts are currently included in the number of simulated users and it should be noted that these scripts must be executed in quantity 1.
As a performance engineer, I always have to explain during the evaluation why some of the scripts are not allowed to be included in the load generation.
In RFSwarm there should be a clear difference between load generation agents and monitoring agents. For the agents for monitoring certain metrics, the number of users can be left at 1, so that RFSwarm does not have to offer the corresponding input field.
With the implementation of the requirement, it will also become easier to correctly display the current load in terms of concurrent users.