0

Running an 8 GB General Purpose v1 Ubuntu / NGINX server on Rackspace and received notification average load was reaching near its peak. Ram usage is also high.

We have multiple sites running on this server so trying to track down if one site is receiving a lot of traffic.

/etc/nginx/nginx.con has access logs set to write to /var/log/nginx/access.log.gz which shows limited traffic at the high load time. Similarly each site has its own access log and have checked those which doen't reveal much either.

Have also checked Google Analytics for our main sites but they don't show much either.

A bit stumped and wondering what else may help uncover what is causing these high server loads and ram usage.

Could a process management issue or memory leak cause this kind of thing?

// edit top output:

PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                                         
 8172 www-data  20   0  477832 124308  40152 S   8.0  1.5   0:16.40 php5-fpm                                                                                                        
31566 netdata   39  19   12812   1656   1272 S   2.3  0.0   0:02.42 tc-qos-helper.s                                                                                                 
 8165 www-data  20   0  463812 107076  37276 S   2.0  1.3   0:10.92 php5-fpm                                                                                                        
 9151 www-data  20   0  456208 102032  39484 S   2.0  1.2   0:10.36 php5-fpm                                                                                                        
 2102 root      20   0  343064  20592   1396 S   1.7  0.3   1729:48 rackspace-monit                                                                                                 
19817 netdata   39  19   14772   3704   1364 S   1.3  0.0   2:15.31 charts.d.plugin                                                                                                 
32098 netdata   39  19   17732   3836    828 S   1.3  0.0   2:38.60 apps.plugin                                                                                                     
 9221 www-data  20   0  464504 106496  35652 S   0.7  1.3   0:08.36 php5-fpm                                                                                                        
13380 ystouch+  20   0  945388  25156   5324 S   0.3  0.3  63:50.95 PM2 v0.15.7: Go                                                                                                 
    1 root      20   0   37204   6328   1324 S   0.0  0.1   2:08.12 init       
  • Please run top / atop and edit your question to show the output. When you run top capture the contents, then hit "M" (or "m") to sort by Memory used, then capture it again. This will tell you what processes are using memory and CPU right now. Do it again in a period of high load if you can. – Tim Feb 28 '17 at 3:03
  • @Tim have updated my original post with top output. load is currently fine so I suspect we will need to run it when load average is high again. – Stretch0 Feb 28 '17 at 3:52
0

You haven't given the information requested (top sorted by memory, atop) but from what I can see something that looks like a rackspace monitoring agent has taken 1729 minutes of CPU time, compared with 63 minutes by what looks like a GO application called PM2.

I would be looking at why the monitoring is taking so much CPU time. Rackspace have install instructions here, but this one might be worth asking their support about.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.