Skip to content

How to properly scale SER?

organiccastleorganiccastle Germany
in GSA Search Engine Ranker
How do I properly scale SER, when is it time for another instance running?

I have learned how to monitor the VPS, proxy usage, Xevil, Xevil proxies but not yet about scaling SER itself.

Obviously, I want to run as many projects as possible on a single instance but also don't want to loose opportunities by not upgrading and spreading the instance over too many projects.
 
What are the indicators telling me it is time to upgrade with another VPS + SER license?
Thanked by 1Deeeeeeee

Comments

  • HunarHunar
    I've noticed anything over 30 projects total seems to be the sweet spot.  Anything else, everything gets spread to thin and you'd have to run it 3x longer to get similar results as just keeping it at 30 projects total.  
    Thanked by 2organiccastle Deeeeeeee
  • rastarrrastarr Thailand
    organiccastle said:
    How do I properly scale SER, when is it time for another instance running?

    I have learned how to monitor the VPS, proxy usage, Xevil, Xevil proxies but not yet about scaling SER itself.

    Obviously, I want to run as many projects as possible on a single instance but also don't want to loose opportunities by not upgrading and spreading the instance over too many projects.
     
    What are the indicators telling me it is time to upgrade with another VPS + SER license?
    A lot depends on the volume of spintext and variable usage etc etc since there's so so much you can shove into SER
    For me, I run 3 Hertzer VPS currently. 2 running SER and the other running XEvil in web server mode that those other 2 VPS access via a tailscale tunnel.

    My 2 SER VPSs run at 1600 threads. 80 projects in the mix, 10 projects run 24x7 and those remaining other projects run within the Project Scheduler, rotating 20 projects every 30 minutes. Took me a little while for tuning but now every runs pretty sweet, with 400-500 VpM on each VPS
    Thanked by 1organiccastle
  • MaximeeMaximee
    edited January 4
    Can someone tell me how to access and set-up the Project Scheduler, in SER? Thanks in advance...
  • SvenSven www.GSA-Online.de
    @Maximee click the down arrow from the start/stop button.
    Thanked by 1Maximee
  • MaximeeMaximee
    Thank you Sven, that's much appreciated  :)
  • malcommalcom
    edited January 4
    rastarr said:
    organiccastle said:
    How do I properly scale SER, when is it time for another instance running?

    I have learned how to monitor the VPS, proxy usage, Xevil, Xevil proxies but not yet about scaling SER itself.

    Obviously, I want to run as many projects as possible on a single instance but also don't want to loose opportunities by not upgrading and spreading the instance over too many projects.
     
    What are the indicators telling me it is time to upgrade with another VPS + SER license?
    A lot depends on the volume of spintext and variable usage etc etc since there's so so much you can shove into SER
    For me, I run 3 Hertzer VPS currently. 2 running SER and the other running XEvil in web server mode that those other 2 VPS access via a tailscale tunnel.

    My 2 SER VPSs run at 1600 threads. 80 projects in the mix, 10 projects run 24x7 and those remaining other projects run within the Project Scheduler, rotating 20 projects every 30 minutes. Took me a little while for tuning but now every runs pretty sweet, with 400-500 VpM on each VPS
    Hi Rastarr

    Can you please elaborate on this one? "2 VPS access via a tailscale tunnel"

    What does that mean? 

    I intend to set up my GSA exactly like this, so I need to determine how to install Xevil on a different machine.

    Edit: I did some research and watched a few videos to learn about Tailscale and how it works. It's absolutely fantastic. From what I understand, you can set it up on different VPSs and create a network. So far, so good. What's the next step after that?

    Do you get the IP address of the Xevil machine from the Tailscale dashboard and then use this IP on other VPS?

    What IP address do you use on the Xevil server to configure Xevil?
  • rastarrrastarr Thailand
    malcom said:
    rastarr said:
    organiccastle said:
    How do I properly scale SER, when is it time for another instance running?

    I have learned how to monitor the VPS, proxy usage, Xevil, Xevil proxies but not yet about scaling SER itself.

    Obviously, I want to run as many projects as possible on a single instance but also don't want to loose opportunities by not upgrading and spreading the instance over too many projects.
     
    What are the indicators telling me it is time to upgrade with another VPS + SER license?
    A lot depends on the volume of spintext and variable usage etc etc since there's so so much you can shove into SER
    For me, I run 3 Hertzer VPS currently. 2 running SER and the other running XEvil in web server mode that those other 2 VPS access via a tailscale tunnel.

    My 2 SER VPSs run at 1600 threads. 80 projects in the mix, 10 projects run 24x7 and those remaining other projects run within the Project Scheduler, rotating 20 projects every 30 minutes. Took me a little while for tuning but now every runs pretty sweet, with 400-500 VpM on each VPS
    Hi Rastarr

    Can you please elaborate on this one? "2 VPS access via a tailscale tunnel"

    What does that mean? 

    I intend to set up my GSA exactly like this, so I need to determine how to install Xevil on a different machine.

    Edit: I did some research and watched a few videos to learn about Tailscale and how it works. It's absolutely fantastic. From what I understand, you can set it up on different VPSs and create a network. So far, so good. What's the next step after that?

    Do you get the IP address of the Xevil machine from the Tailscale dashboard and then use this IP on other VPS?

    What IP address do you use on the Xevil server to configure Xevil?
    Yes, you grab the Tailscale IP for the XEvil machine and use that for both the Xevil webserver setting as well as using that same IP on the other VPSes
  • malcommalcom
    edited January 5
    @rastarr

    1) In this scenario, Xevil IP will remain as 127.0.0.1, correct? Since they are connected via Tailscale, there is no need to change it to the server IP.

    2) Similarly, we can use the server's IP address of the Xevil machine in the Xevil settings and use the same IP address in the GSA settings. I believe this method will also work without Tailscale.

    So, what advantages does this method have over the traditional method?
  • rastarrrastarr Thailand
    malcom said:
    @rastarr

    1) In this scenario, Xevil IP will remain as 127.0.0.1, correct? Since they are connected via Tailscale, there is no need to change it to the server IP.

    2) Similarly, we can use the server's IP address of the Xevil machine in the Xevil settings and use the same IP address in the GSA settings. I believe this method will also work without Tailscale.

    So, what advantages does this method have over the traditional method?
    That is what I am using, is all. 
  • FreshairforlifeFreshairforlife earth
    malcom said:
    @rastarr

    1) In this scenario, Xevil IP will remain as 127.0.0.1, correct? Since they are connected via Tailscale, there is no need to change it to the server IP.

    2) Similarly, we can use the server's IP address of the Xevil machine in the Xevil settings and use the same IP address in the GSA settings. I believe this method will also work without Tailscale.

    So, what advantages does this method have over the traditional method?
    With tailscale  you can can have access to other vps while xevil just use as a webserver. What I do is setting up gsa cb as a webserver and then add xevil as secondary.  My setup is very similar as one vps dedicated only for captcha blasts. 
  • KushphktKushphkt TMobile
    @organiccastle I just upgraded to xrummer Ai and the performance wise you can't beat xrummer as their operators and engines are beyond the level. So don't try to compare gsa ser with xrummer at all .I didn't get time for new upgrade as I just did but as a former user of xrummer  it's so unfair to compare those. 

    Whenever you see often ram issues, resources issues no matter what and how you regulate scheduler it's time for a new gsa intemce.

    I am agree with what @rastarr said about spintax part. Also use spinfile macros when loading articles and such data it lowers the load.

    @rastarr how do you get 400vpm.. I am getting ok vpm around 50 - 60 at 400 treads to  sometimes 800 but have not experienced such.. I purchae few link lists at a time and combine them and use after sometime threads going lower and lower unless I load new links to it no matter what. @rastarr may I know what proxies provider are using for that?



  • organiccastleorganiccastle Germany
    @Kushphkt

    Thanks. I've not used any Spintax in most projects since the OpenAI API was released in 2021. I'm loading unique articles into SER and now just deployed my own API through the %file-% macro to generate relevant blog comments rather than using pre-written and somehow niche relevant comments. (The mistral API is giving good results and is cheaper than OpenAI. Success rate on blog comments being published way better.) Load is therefore not on the SER VPS for spintax or any other macro processing.

    I experienced the figure of 30 projects @Hunar kindly mentioned to be realistic. Attempts to run more projects through the project scheduler did not show better results.
  • KushphktKushphkt TMobile
    organiccastle said:
    @Kushphkt

    Thanks. I've not used any Spintax in most projects since the OpenAI API was released in 2021. I'm loading unique articles into SER and now just deployed my own API through the %file-% macro to generate relevant blog comments rather than using pre-written and somehow niche relevant comments. (The mistral API is giving good results and is cheaper than OpenAI. Success rate on blog comments being published way better.) Load is therefore not on the SER VPS for spintax or any other macro processing.

    I experienced the figure of 30 projects @Hunar kindly mentioned to be realistic. Attempts to run more projects through the project scheduler did not show better results.
    20 projects or 30 is a sweet spot. Mistral is great specially if you want uncensored. Btw I am a open ai fan and I use all fields with GSA CG but still spin big part of my projects. Spinnerwriter does the trick.


  • royalmiceroyalmice WEBSITE: ---> https://asiavirtualsolutions.com | SKYPE:---> asiavirtualsolutions
    organiccastle said:
    How do I properly scale SER, when is it time for another instance running?

    I have learned how to monitor the VPS, proxy usage, Xevil, Xevil proxies but not yet about scaling SER itself.

    Obviously, I want to run as many projects as possible on a single instance but also don't want to loose opportunities by not upgrading and spreading the instance over too many projects.
     
    What are the indicators telling me it is time to upgrade with another VPS + SER license?

    To start with get a decent vps  with at least 6GB of Ram ( +/- 3gb for windows and 3 GB for GSA) get a decent cpu - If you need a good vps package then visit my site and look for the GSA ser VPS package

    But you can run more than a 1500 projects on one machine thats not a problem, you just need to play with the status and the scheduler.

    The most important thing is to monitor your CPU \ Memory \ Threads \No. of projects to run in scheduler

    Here is a quick video showing you how i run 1,674 projects all at the same time.

     https://www.loom.com/share/c12271055588419eb69a28e99ffbbcee?sid=cfd425ef-8032-401b-bf85-b6d70a030479

    GSA scheduler - 14 April 2024 - Watch Video


    Thanked by 1organiccastle
Sign In or Register to comment.