3
Resource counter_cpu_share_used red alert on environment 170 dominio.com current value: 89 soft limit: 85 hard limit: 95
Sometimes in the resource log of this red warning, what worries me, the question is.. How can I debug this in Powerpanel? I tried to see something in the processes that are being executed, but I can’t understand how this can help me.
Server is VPS.
I will leave the printado System Proccess, in order to facilitate the understanding of my current scenario.
3
Good night,
I have no experience with power panel, but I have with other similar management software and the post has little information. But I’ve studied some material from the net and I’ll try to help you.
For starters, here there is a PDF (in English), with administration information of this system.
On page 62 (Monitoring Operations and Viewing Logs) talks a little about the display and configuration of these logs.
This is happening because the processing has reached a preset limit in the parameter counter_cpu_share_used. To debug, you have to analyze the list of processes in Container Services Dashboard by going to the link System Processes and see which processes are consuming the most resources, through the TIME column, you see which processes are using the most CPU time. Then you need to study whether you should change the active application/service, restricting use if possible, or whether you should increase your VPS CPU.
From what I’ve seen of the images you posted, it’s a quantitative indicator, which means it only counts the number of active processes. This will only be a problem if the % of the CPU is high.
It is common for apache to make processes available for each connection. Here there is a detailed explanation of this and how to limit the processes (in English).
Follow a passage:
For example, Apache needs 1000 processes to Handle 1000 concurrently Connected clients, or Connections.
translating:
For example, Apache needs 1000 processes to handle 1000 clients connected simultaneously, or connections.
If cpu time (in %) is not auto, just adjust the parameter counter_cpu_share_used, for a more auto value, otherwise you will have to study an action for that service, whether it be improving processing or restricting simultaneous users/connections.
Any questions, I’m available.
Browser other questions tagged powerpanel
You are not signed in. Login or sign up in order to post.
Good! I will process this information and come back after here to see what I could understand, thank you!
– Ale
Printei ali (I put in the post) the last part of the processes in the System Proccess session, what worries me is the Command that seems repeated, with high team(bigger than the other processes,. This Command repeats over and over again:
/usr/local/apache/bin/httpd -k start -DSSL– Ale
I changed the answer with a new reference. What is the % of cpu time? If it is not high, you can change the parameter, otherwise you will have to analyze more deeply what is generating auto consumption.
– Eric Silva
I do not understand when you say: "What is the % of the cpu time",.. What I’m surprised that the CPU used says 49%, when the parameter
cpu_share_usedis almost bursting(near 95 when the limit is 95). Thanks for the attention, and the weather of course.. thanks!– Ale
Dispose. As to the doubt, because the counter_cpu_share_used is counting the number of processes in use while the 49% of CPU usage is the time the CPU consumes computing in percentage. This means that 51% of the time it is idle, so I asked "What is the % of the cpu time". I think the solution would only change the log parameter and put something around 130 processes as critical, which by the data you told me, would represent on average 70% of the CPU usage and not more unnecessary alerts in the log.
– Eric Silva
Eric, I’ve been looking for how to change the parameter
counter_cpu_share_used, but without success. I believe that changing this parameter my server will drop (I realized that it drops when the current value exceeds the limit(95)). How can I change this value? Even if I fall again at least I would no longer display these alerts, which already helps. Thank you.– Ale
Alexandre, I found here a detailed description of the problem and how to solve it. (In English). Here also has a good source of information, to make a study in the processes and see if it is hack, defect in the application or if it is normal. It seems to me that the hole is lower than we were thinking. This warning, from what it says in this article, is really the use of the cpu and not a simple process number counter. Which means your container is really overloaded.
– Eric Silva
Just for the record, I’ve decided to switch to a dedicated plan, and I’m going to start working with memcached to make the server more performative, Eric Hanks, thank you for your persistence in my case, even more.
– Ale