Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


A tool to manage all your idlers?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

A tool to manage all your idlers?

angeliusangelius Member
edited June 2019 in General

Hey dear LET community,

I'm now the proud owner of few idlers vm from around the world. Each one is with a different provider (trying to create a super mesh).

I wonder what kind of tool you guys are using to monitor/control all your idlers.

I'm searching for something like ssh+ansible+replicated storage+global monitoring/alerts all that in one panel to master them all :)

Do something like this already exist?
Like you install the master node, you connect all the other nodes from arround the world to it. Then you can manage all your vm in one place.

If yes, please name it and I will probably give it a try :)

If not, I might consider to develop one (based on my favorite lightning fast stack: node.js cluster+redis). Would you be interested?

CloudMaster: manage all your idlers
  1. Is it the kind of tool you would like to use?36 votes
    1. I'm interested!! I have tons of idlers to manage
      63.89%
    2. Not interested
      36.11%

Comments

  • rskrsk Member, Patron Provider

    Check it out. https://github.com/myrsk/vmdash

    Not as feature rich, but interacts with providers via api.

    If you have cool features you want to add, we can work together :smiley:

  • williewillie Member

    I don't see any need for this. I'm too paranoid to leave ssh private keys on any of my servers, so I run ansible from my laptop. I don't have any serious monitoring for my idlers other than what the host supplies, since who cares if an idler goes offline for a few minutes? But if I wanted monitoring I could set it up myself or use one of the existing monitoring services.

    Thanked by 2angelius vimalware
  • @rsk said:
    Check it out. https://github.com/myrsk/vmdash

    Not as feature rich, but interacts with providers via api.

    If you have cool features you want to add, we can work together :smiley:

    Wow thanks for this, I will definitely take a look and try it :)
    Would you mind to share few screenshots?

    @willie said:
    I don't see any need for this. I'm too paranoid to leave ssh private keys on any of my servers, so I run ansible from my laptop. I don't have any serious monitoring for my idlers other than what the host supplies, since who cares if an idler goes offline for a few minutes? But if I wanted monitoring I could set it up myself or use one of the existing monitoring services.

    Interesting point of view (security wise), but in that case why not just install the Master node in your laptop instead of over internet? This way the SSH private key will exist only on your laptop, that would fix your concern and still provide a panel monitor/control all your VMs

    I guess this kind of system is needed when you have more than 5-10 VM to manage

  • williewillie Member

    My laptop isn't always online and I don't feel like I need to run monitoring from it. If I was to monitor seriously I'd use multiple VM's from different geolocations since lots of the outages I experience are route specific. I rarely see VM's actually crash. It's usually network issues that get fixed without my intervention, so getting alerts about them would only annoy me. ;-)

    I do think I have more than 10 VM's but probably less than 20 these days. It used to be more.

  • @willie said:
    My laptop isn't always online and I don't feel like I need to run monitoring from it.

    Monitoring can be done by each node but only accessible securely through the panel. So even if your laptop is offline monitoring will continue and alerts still sent by email

    What about monitoring disk space, nginx, mysql, redis, postgix.... Manage versions, security upgrades, penTests, ...

    I have something like 30 VM and it's a pain to setup monitoring / control for each of them

    I guess the need highly depends on what you are doing with your VMs

    Personally I have a lot of everything, email servers, web (nginx, apache), db (mysql, postgres, redis), load balancers, build servers for CI/CD ...

  • I just use spreadsheet with color codes (severity) to remind me how much idling I'm currently up to.

    Thanked by 2angelius plumberg
  • williewillie Member
    edited June 2019

    angelius said: Personally I have a lot of everything

    You're not describing idlers then. I see you're still pretty new here and probably don't have a REAL idle fleet yet ;-).

    More seriously I try to build stuff for low attention, which means I use cron scripts instead of daemons when I can, append collected info to disk files and process it later instead of having online databases, etc. I have a few small web sites on shared hosting, and only run one http server which has been very solid. I check it once in a while and maybe I should set up something to check it daily, but that's about the one thing, and it's on a shell account on someone else's colo box, rather than a VM.

    For the other stuff, regarding email alerts: do not want!!! I'm used to getting 100s of those a day for stuff at work. I don't want to spend time or aggravation looking at anything like that unless someone is paying me to. Otherwise if something is down I'll notice when I try to use it.

    Thanked by 2angelius vimalware
  • @vimalware said:
    I just use spreadsheet with color codes (severity) to remind me how much idling I'm currently up to.

    I do the same for now ... But it's getting complex and error prone with more than 30 VM ...

  • Just idle them and use as needed. No need to monitor idlers lol

  • defaultdefault Veteran

    If I would monitor all my idling servers, it means I am the one who does not idle anymore, because I need to keep watching and analyze if the servers are up or not, or why those are not idling properly. Therefore, the servers are idling well and fine, while I have to work and see if they idle as intended.

    Meanwhile, the server (or service) that monitors all these servers is not idling either, working constantly to monitor and maybe even trigger alarms which defeat the whole purpose of idling; not to mention you get one less server to idle.

    Whenever an idling server goes offline I will have to go and complain on LET (& HostBalls & vpsBoard) about how a provider failed to deliver my expectations, while the server was used for production idling and I lost all data.

    Sometimes is better to not ask myself too many questions... and let the idling server do it's thing.

  • Production idling servers

  • uptimeuptime Member

    if a server idles in a datacenter, and no monitor is there to display it ...

    (idling intensifies)

  • rskrsk Member, Patron Provider

    @angelius said:

    @rsk said:
    Check it out. https://github.com/myrsk/vmdash

    Not as feature rich, but interacts with providers via api.

    If you have cool features you want to add, we can work together :smiley:

    Wow thanks for this, I will definitely take a look and try it :)
    Would you mind to share few screenshots?

    @willie said:
    I don't see any need for this. I'm too paranoid to leave ssh private keys on any of my servers, so I run ansible from my laptop. I don't have any serious monitoring for my idlers other than what the host supplies, since who cares if an idler goes offline for a few minutes? But if I wanted monitoring I could set it up myself or use one of the existing monitoring services.

    Interesting point of view (security wise), but in that case why not just install the Master node in your laptop instead of over internet? This way the SSH private key will exist only on your laptop, that would fix your concern and still provide a panel monitor/control all your VMs

    I guess this kind of system is needed when you have more than 5-10 VM to manage

    Sure, check out this thread for screenshots

    https://www.lowendtalk.com/discussion/147713/opensource-global-server-manager-in-laravel/p1

    Thanked by 2angelius vedran
  • JanevskiJanevski Member
    edited June 2019

    There needs to be shut down and power on, on demand your idling VPS servers initiative.
    It's even better than idling, you fully support the providers and use only when and if needed.
    Just like communism, but, this time by the users, for the providers, with love.

    There should be an automatic shutdown script if there is no traffic or load above X after a certain period of time, on a server marked for on demand usage.

    Thanked by 2uptime plumberg
  • I had to create an account just to thank you guys for this thread.

    I have a question though. Is it better to have a lot of small idlers, or just a few really big ones. On one hand small idlers will idle more ip addresses, but big idlers will have a larger percentage of actual idle resources.

    it's a damn riddle wrapped inside of an enigma.

    Thanks.

    Thanked by 1angelius
  • @captainwasabi said:
    I had to create an account just to thank you guys for this thread.

    I have a question though. Is it better to have a lot of small idlers, or just a few really big ones. On one hand small idlers will idle more ip addresses, but big idlers will have a larger percentage of actual idle resources.

    it's a damn riddle wrapped inside of an enigma.

    Thanks.

    I once sat down and started making an algorithm to score idlers based on resources. Then I realized I didn't actually give a fuck after all.

    Thanked by 1angelius
Sign In or Register to comment.