Thanks for the information, very useful. Looking at the numbers you 
provided below, it looks like the changes you have made are doing the trick 
- load is well under 4 and memory is around 80% though you may need to add 
more memory in the future.

On Sunday, May 10, 2015 at 6:54:53 PM UTC+2, BKeep wrote:
>
> Currently, the log volume is about 1200 logs per second. I have an ES 
> cluster setup with two data storage nodes and an ES node that functions as 
> master only.. The master only node is installed on the same box as graylog, 
> graylog web, and mongodb. Adding the additional 2 vCPU's took care of the 
> problem for now but we plan to add logging for another 200+ devices so am 
> not sure where that leaves us. The current master node has 4 vCPU's, which 
> are XEON 2.7Ghz procs and 12 GB of RAM allocated to it. CPU load is -- load 
> average: 1.92, 2.08, 1.96 and memory usage is -- Mem:  12198260k total,  
> 9416592k used,  2781668k free,   258876k buffers. I changed the HEAP size 
> to -Xms4g -Xmx4g from the original setting of 1g.
>
> On Saturday, May 9, 2015 at 5:42:25 AM UTC-5, Matt Maloney wrote:
>>
>> How many messages are you ingesting per second? Zi's recommendation is 
>> right, you should consider running ES separately from Graylog. Moreover, 
>> you will want a load balancer in front of the Graylog servers. I imagine 
>> with the firewall logs being added into the mix, things have gotten a lot 
>> chattier hence the load issue. 
>>
>> On Friday, May 8, 2015 at 3:31:27 AM UTC+2, BKeep wrote:
>>>
>>> We just setup our ASA to send logs into graylog and It appears the 
>>> graylog setup is not keeping up with the incoming log data. The server 
>>> currently has 8 GB of RAM and 2vCPU's. Up until the addition of the ASA, 
>>> that setup has been keeping up pretty well. However, it looks like i am 
>>> going to have to make some adjustments. Prior to adding the ASA logs we 
>>> were logging for about 130 servers. My question is what would be the next 
>>> steps? Obviously add some Memory resources and another couple of vCPU's but 
>>> once the ram is added are there any other adjustments to be made on the 
>>> server itself?
>>>
>>> Thanks in advance..
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to