Kaseya Community

How often should an Agent check-in?

This question has suggested answer(s)

Hi,

 

I am editing all the templates that we have and creating a few new ones. I'm just wondering what the best check-in period is? Currently we are using 1 minute, but we manage over 2000 machines and we think that the added traffic every minute is causing our network and server to run slower than is normal.

 

I was thinking maybe every 2 or 5 minutes? Would this cause any problems? Short or long term?

 

Thanks in advance,

Rebecca.



[edited by: MissRAE at 9:07 PM (GMT -7) on 10-18-2010] changed from discussion to question
All Replies
  • Hi.

    I´m also intrested in what other peoples suggest here.

    We have around 4500 agents at the moment but will reach 7500 in the next year.

    right now we have 1 minute.

  • I've put servers on 90 seconds and workstations on 150. Absolutely  no scientific basis to it but a minute felt like it was too short and that touch more traffic. I guess it depends how often you do something that needs an "interactive" response such as kicking off a live connect / remote control session or an agent procedure that's going to do something for a user who is on the phone right now...

    I may get some pressure to change the workstation period once our service desk starts using Kaseya in real anger but it seems to do fine just at the moment.

    Just my 2p worth.

    Graham.

  • We are managing ~750 Agents, but with a very thin bandwith mostly in our clients side. So we are using 30sec for servers and 3 min for others. Other important factor could be BW limit. In our case, we use 15KB/s and 5KB/s for Servers and workstations respectivly.

    So far we could manage all the base very well and without complaints under stipulated SLAs.

    Also keep watch on distribution schedules, Patchscan & other BW or transactions heavy procedures (like merging agents).

    JRodino

  • We use 1 min for servers and 4 mins for workstations.

    The thing to remember with check-ins is that it will try and keep the tunnel open between check-ins. This means when you fire up RC it will often connect before the next check-in unless there are problems with connectivity.

    We have been using 4 mins here for years and a lot of the time RC will connect in a minute or less. I've often found when the RC takes more than that to connect it's because something is wrong or there is another script running before my RC script (still on V5 here).

  • I have my K server in a data center (lots of bandwidth). I also have 2500 agents checking in. I have the K server defaults for all devices (30 secs I believe). FYI - My bandwidth consumption is low.

  • Hi David,

    This is an interesting response. My concern over this has not been hte bandwidth so much as the demand on the Kaseya server. We found when we had an older server it slowed the system down too much.

    I would be curious to see how the system handles it now that we have upgraded.

  • It doesn't seem to have that much of an impact on my K2 server (nor my K1 server). My tech report that it is working at normal speed and they do not experience slowness issues.

  • I am not sure that the check-in really has any affect anymore.  In the old days, the Kaseya agent used to check-in.  You would have to wait for a script to run or remote control, etc.

    These days the agents are in a "constant connection".  They are always checked in and ready to act.

    I have wondered many times why the check-in feature is still available.  I don't think it does anything anymore.  Maybe someone on Paul's team can chime in here and clarify.

  • well i just changed all my agents to 30 seconds. Didn't see a big performance impact but RC is firing up a bit quicker at times. a change to 30 seconds is one I would definitely recommend.

  • Question:

    Priority:Description:Hi,

    I was just wondering if someone could please tell me how the check-in period works and what it affects?

    Reading the administrators guide has left me a little confused, from what I understand, it is used for if a connection gets dropped. Yet there is a column you can select that is called "Last Check-in" which from the above understanding means that each agent is dropping a connection very regularly. I thought it just meant how often an agent checked in with the Kaseya server?

    If my second definition is correct, would there be a large problem if I changed it from 1 minute to 2 minutes in Agent -> Configure Agents ->Check-in Control?

    Cheers,

    Rebecca.

    Answer:

    Rebecca,


    the check in control is used for when an agent checks into the server, this is done at regular intervals (30 seconds is the default) so that when you send a command to the machine you dont have to wait several hours for it process.

    a good example is monitoring, if you set the check in time every hour and the agent goes down just after the last check in you would not be informed that the agent was down until it checks in again.

    The last check-in column simply specifies when the last check in was, this is usually only used to check when a machine went offline as you can see almost to the exact time when it last had any communication with the server.

    There is no problem changing it to 1 minute or 2 minutes.

    regards

    Kaseya Support

    ____

    Hi,

    So just to make sure I understand corectly, the Agent keeps in constant contact with the Kserver and just does a quick check-in every 1minute (this is what we are set to)? Changing the check-in time won't affect how long it takes to connect to live connect? Do you feel that changing the workstation check-in period to 2 minutes will help lower traffic? I posted a question on the forum about this same topic, and people have come back saying things like "servers have 30sec periods and workstations 3 minutes and there have been no complaints", would a bigger period like the 3 minutes cause a reduction in usuability when it comes to remote control and audits etc.

    Is there anyway to limit the bandwidth usage of each agent? We are experiencing a fair bit of traffic at the moment and I am trying to fix the problem.

    Also, do you mind if I post this ticket on the forum, a couple of people have shown interest in the topic?

    Cheers,
    Rebecca.

    ___

    Rebecca,


    correct.
    Checkin will not affect live connect. Changing work stations to two minutes is fine, it will help lower traffic but the agent traffic is minimal I would have to dig out the calculation tomorrow when I'm in the office again, I'll be able to tell you exactly what each agent uses - on average of course.

    The usability wil be the same, it will only affect the run now functionality as it can then happen anywhere within that 3 minute window (depending on last check-in) scheduled items like audits will still run as per the schedule.

    You can limit the bandwitdh of each agent this is done under the check-in control page where you specify the address/frequency and bandwidth.

    regards

    Kaseya Support

    ___

     Ok, this looks like it is worth trying. With the bandwidth, will we notice a reduction in stress on the server? or our network? because that is the aim. 


    Is it ok if I post this ticket on the forums? 
    _____
    Rebecca,

     

    Yes you may post this on the forum.

     

    By reducing stress I assume you mean the Kserver? I will attach our best practise performance guide for you, this goes through some items like the amount of agent vs RAM and CPU etc.

    http://dl.kaseya.com/download/support/vsa/performance/bestpractices.pdf

     

    regards

     

    Kaseya Support

  • I donot think you should worry much on this section. We are using below settigns for years and never had an issue:

    checkin period - 30 seconds (both for server/workstations)

    bw throttle - disabled