Global cluster usage overview on a per satellite basis
Posted: Mon May 09, 2022 3:32 pm
Hello,
On Eve-NG Cluster Pro, the "System > System status" menu gives the total number of qemu nodes started, another menu "system > cluster management" shows the global load of each cluster member and eventually, the "management > Node Management" and "Management > Lab management" give a list of all nodes, their state and cpu/memory consumption.
Would it be possible to have a menu that gives an overview like this :
- vCPU/RAM/hosts distribution per user and per satellite (vs the system capacity)
- a global table that aggregates everything (optionnal)
The aim is to let user see which satellite is more or less used and make sure their lab are properly distributed among the cluster. This would ease not to overload one satellite member.
I included the satellite capacity (I guess you can easily get it from the OS : number of vCPU / amount of RAM)
In the example below, we can see that satel-04 is not used at all although satel-03 is overloaded (100vCPU used when it only has 96 for real...)
On Eve-NG Cluster Pro, the "System > System status" menu gives the total number of qemu nodes started, another menu "system > cluster management" shows the global load of each cluster member and eventually, the "management > Node Management" and "Management > Lab management" give a list of all nodes, their state and cpu/memory consumption.
Would it be possible to have a menu that gives an overview like this :
- vCPU/RAM/hosts distribution per user and per satellite (vs the system capacity)
- a global table that aggregates everything (optionnal)
The aim is to let user see which satellite is more or less used and make sure their lab are properly distributed among the cluster. This would ease not to overload one satellite member.
I included the satellite capacity (I guess you can easily get it from the OS : number of vCPU / amount of RAM)
In the example below, we can see that satel-04 is not used at all although satel-03 is overloaded (100vCPU used when it only has 96 for real...)
Code: Select all
user | CLUSTER | satel-01 | satel-02 | satel-03 | satel-04 |
total | 143 hosts / 262 vCPU / 522240 Mb | 52 hosts / 83 vCPU / 172032 Mb | 41 hosts / 79 vCPU / 145408 Mb | 50 hosts / 100 vCPU / 204800 Mb |
2 | 107 hosts / 206 vCPU / 428032 Mb | 39 hosts / 70 vCPU / 145408 Mb | 26 hosts / 52 vCPU / 102400 Mb | 42 hosts / 84 vCPU / 180224 Mb |
1 | 4 hosts / 4 vCPU / 8192 Mb | 1 hosts / 1 vCPU / 2048 Mb | 3 hosts / 3 vCPU / 6144 Mb | |
3 | 32 hosts / 52 vCPU / 86016 Mb | 12 hosts / 12 vCPU / 24576 Mb | 12 hosts / 24 vCPU / 36864 Mb | 8 hosts / 16 vCPU / 24576 Mb |
HOSTS distribution
user | cluster | satel-01 | satel-02 | satel-03 | satel-04 |
total | 143 hosts | 52 hosts | 41 hosts | 50 hosts |
2 | 107 hosts | 39 hosts | 26 hosts | 42 hosts |
1 | 4 hosts | 1 hosts | 3 hosts | |
3 | 32 hosts | 12 hosts | 12 hosts | 8 hosts |
vCPU distribution
user | cluster | satel-01/112 | satel-02/112 | satel-03/96 | satel-04/112 |
total | 262 vCPU | 83 vCPU | 79 vCPU | 100 vCPU |
2 | 206 vCPU | 70 vCPU | 52 vCPU | 84 vCPU |
1 | 4 vCPU | 1 vCPU | 3 vCPU | |
3 | 52 vCPU | 12 vCPU | 24 vCPU | 16 vCPU |
RAM distribution
user | cluster | satel-01/385574 | satel-02/385574 | satel-03/257395 | satel-04/385574 |
total | 522240 Mb | 172032 Mb | 145408 Mb | 204800 Mb |
2 | 428032 Mb | 145408 Mb | 102400 Mb | 180224 Mb |
1 | 8192 Mb | 2048 Mb | 6144 Mb | |
3 | 86016 Mb | 24576 Mb | 36864 Mb | 24576 Mb |