feature suggestion - Next Proxmox Releases

[starnetwork]

Hi,
First, Thanks to Proxmox team for all the good work!
and also, Congratulations for the 6 years for Proxmox VE project

few suggestions for next releases:
1. Improving the backups, both add more flexible backups Scheduling (monthly backup?) and differential backup (Avialable as 3d party plugin but should be in core and available on any new version)
2. GUI Update - Proxmox GUI moved from v1x to v2x big upgrade but now it's the time to the next step, all other Competitors projects Adopted the "Amazon Style" GUI, in my opinion Proxmox should be there also
3. Resources alerts - add option to send alert when VMs / Node CPU / Memory is over loaded or Disk drive is for example 95% Full

Again, thanks for everything And I wish all the best for staff and community.

Best Regards,
Star Network.

 

[mo_]

2. GUI Update - Proxmox GUI moved from v1x to v2x big upgrade but now it's the time to the next step, all other Competitors projects Adopted the "Amazon Style" GUI, in my opinion Proxmox should be there also

Which "competitors" are you speaking of? Because VMware Vsphere hasn't changed its GUI for quite a while. Also why would you ask for a UI downgrade?

3. Resources alerts - add option to send alert when VMs / Node CPU / Memory is over loaded or Disk drive is for example 95% Full

That's what icinga (just to name one) is for. Proxmox is a virtualisation wrapper/clustering solution and as such this (to me anyways) seems out of scope. Why try to reinvent the wheel on a field that has lots of possible solutions already

 

[starnetwork]

Hi,
2. talking about openstack, cloudstack, openshift, and even other same solutions like solusvm
3. am not talking about downtime alerts but resources status, in my Opinion it's should be a Proxmox module.

All the best,
Star Network.

 

[cesarpk]

Hi to all

Talking about of alarms into PVE, I agree with what starnetwork says, VMware and Citrix have these options incorporated, these type of features do the difference between a product with a full solution or not, and several people that not have much experience in Linux will want to have this solution in a only product and not to have that install and configure a VM or CT for separated for get these additional features (they sometimes nor know for where should start for get these features nor to who ask, so that may be that PVE will be discarded quickly).

As a example, i like and want to see in the PVE GUI the resources monitor, by each Core of CPU, each Network (bonding or simple) and with report more accurate in time of analysis (in Citrix free version i can see it for each 5 seconds if i want)

Anyway, PVE is a product open source, very great in options, and have another features interesting that the competitors don't have in his free versions.

Best regards
Cesar

Re Edited: Also, these competitors of hypervisors can analize if a VM have some resource saturated or inclusive when is soon for saturate, and can move the VM(s) automatically to another Host(s) that have the higher free resources, also can give us the alarms of actions made by several ways (console center, mail, logs, etc.)

 

[spirit]

Re Edited: Also, these competitors of hypervisors can analize if a VM have some resource saturated or inclusive when is soon for saturate, and can move the VM(s) automatically to another Host(s) that have the higher free resources, also can give us the alarms of actions made by several ways (console center, mail, logs, etc.)

Yes, I would like to work on this this year (workload auto balancing)

 

[spirit]

2. GUI Update - Proxmox GUI moved from v1x to v2x big upgrade but now it's the time to the next step, all other Competitors projects Adopted the "Amazon Style" GUI, in my opinion Proxmox should be there also.

whmcs have a nice module for proxmox
http://www.whmcs.com/
http://www.modulesgarden.com/products/whmcs/proxmox_vps/features

maybe this can help your need ?

 

[e100]

+1 Differential Backups

I like the GUI how it is, being able to do more in the GUI to avoid the CLI would be an improvement.

Resource alerts would be nice, but there are numerous tools to do that so it is not a pressing concern, I use Zabbix.
If they ever implement VM auto balancing I suspect that load alerts would be an easy byproduct of that.
To balance load once must monitor it.

 

[spirit]

+1 Differential Backups

should be ready this year in qemu (I have already seen prelimary patches on qemu mailing list).
Just need to wait

 

[methil]

Hi,
1. Improving the backups, both add more flexible backups Scheduling (monthly backup?) and differential backup (Avialable as 3d party plugin but should be in core and available on any new version)
2. GUI Update - Proxmox GUI moved from v1x to v2x big upgrade but now it's the time to the next step, all other Competitors projects Adopted the "Amazon Style" GUI, in my opinion Proxmox should be there also
3. Resources alerts - add option to send alert when VMs / Node CPU / Memory is over loaded or Disk drive is for example 95% Full

1: Backups need some rework.
- Retention defined per job and not per storage node. Or at least allow job to override the setting.
- Option to name the backup file as we want and need. qemu-vzdump-VMID-DATE.tar.gz is not enough and simply useless. What and who is VM 136 or CT 234?
- Better scheduling, every x min, hour, day, week, month, year. Specific day of week or month. First occurrence of "insert day name" of month. recurring every x time.

2: Gui Update: Rewriting the gui require a lot of time. Time that i prefer they spend on feature like backups.
- More infos displayed per vm would be nice. no graph just text: current usage of each core, ram, swap, IO read/write, same for each network interface.
- More infos displayed for the host, current IO read write, network per interface and bond/vlan.
- Current graph need to poll the usage more frequently. 5 mins polling/average is to slow and inaccurate for CPU and network usage.
- Option to dig usage graph like cacti. Minute / hour / day / week / month / year average. Reporting module with scheduling?

SNMP PROXY : need to open some RFE for this.
- Allow you to connect to any node and get proxied to the right node to get the resources used on the host or by a CT / VM in a cluster.
- This will allow anyone to use the software they want to monitor what going on and trigger any type of alert and/or do any type of usage accounting.