At first I thought “1” represented 100% load, but then I saw it jump above 1, to 1,2 and the graph changed to where “2” was the highest number shown. Guessing this depends on the number of cores, but I’m not sure how many cores this one has, so how do I read it and what is max?
Bonus question, is there a general max number of connections this device can handle? Or a number of connections that is not recommended to exceed?
Can you share with us any screenshots showing that it is jumping above 1? As per Status → Realtime Data → Load, the maximum value is 1. Note that this device is running on 7.05.4 Firmware.
There is no hard limit on how many devices are allowed to be connected. But do note that bandwidth and speeds of the connectivity will vary depending on how many devices are present in the network.
For Wi-Fi connectivity, RUTX50 can handle up to 150 simultaneous users.
Hope this cleared your doubts.
Hello. Here is a screenshot:
For the other question, I was more interested in number of connections, as in TCP and UDP connections (that you can see in realtime data), not number of connected devices. This router does not have wifi enabled, but is connected to a wifi mesh network with perhaps up to 150 devices. I’m assuming it can handle more clients when the AP’s are external?
I see number of connections going to well over 2000 at times and I’m wondering if this might be on the upper end of its ability, but I also understand this depends on various factors such as bandwidth etc.
Here is a screenshot of the connections. It sometimes goes even higher, but for all I know it might well be within the limits:
This graph shows CPU load how it is portrayed in Linux systems.
Same logic applies here since RutOS is based on OpenWRT.
Short explanation about number interpretation:
Interpreting the numbers
Let’s take a hypothetical load average value - 1.73 0.60 7.98 on a single CPU system.
1.73 during the first minute denotes 1.73 runnable processes. That means after optimum traffic the CPU still had 0.73 processes waiting for a chance to load.
In the next five minutes, the load is lesser than the prescribed amount. Only 0.60 processes are slated for running against a capacity of 1.00, thus the CPU was idle 40% of the time.
In the last fifteen minutes, the system was heavily overloaded. 7.98 processes against the required 1.00 were waiting to be executed.
This means that 6.98 processes were on the waiting list when the single CPU server ran once. This is a very high value for an average and hence has to be fixed immediately.
This explanation is taken from Understanding Load Average in Linux where you could find more information about different scenarios.
Thank you. I’m assuming the RUTX50 is single core?
RUTX devices use Quad-core ARM Cortex A7, 717 MHz CPU.
Thank you. Am I understanding it correctly that the number would then have to reach 4 before there is 100% utilization?
That is correct. If the CPU is overloaded and cannot handle all threads, this value could go over 4. However, this is extremely unlikely to happen in real-world usage, and even if it does, the device would simply work slower, but no other side effects should appear.
Thank you. Marked as solved. The reason for my inquiry is that we are actually having an issue where wifi clients sometimes loose connectivity (external APs). Sometimes they loose it under roaming, a few times when connecting (not getting IP). I wanted to know whether it could be load that did it, but seems we’re well within tolerance levels with that then.
This topic was automatically closed after 15 days. New replies are no longer allowed.