[ale] System Load Summary Script?

[Todor Fassl]

I keep a pretty close eye on my systems via nagios.  Jeffery posted that 
article about how the numbers you get when you run uptime came to 
include some measure of disk IO. And there are approximately eleventy 
gazillion blog posts on google explaining how to drill down to find out 
what is wrong if those numbers are high. I just think that by now, there 
ought to be some command someone wrote to automate the process explained 
in those blog posts. This tool "glances" is pretty close.

Anybody familiar with mysqltuner? It queries dozens of mysql parameters 
and prints [Okay] if they are within an acceptable range or [!!!!] if 
not. If I get a problem notice, I google it. Is that really a problem? 
Mysqltuner is pretty much always spot-on.





On 6/26/19 3:55 PM, DJ-Pfulio via Ale wrote:
> I use monitoring.
> 
> Monit or munin or something else to capture everything needed every few
> minutes.  With the graphs, you can quickly see if something changed the
> last week, month, quarter or year.  You already have the data, just need
> to look at it.  If you don't do the monitoring, there isn't anyway to be
> proactive that I know to recognize issues besides whatever ends up in a
> logfile and gets parsed.  By that point, it is way too late.
> 
> Alarming is different than monitoring.  If the tool you use can't be
> tweaked NOT to alarm on system-normal behavior, then you need a better tool.
> 
> None of these things is install and forget. They require configuration
> and tweaks from a professional admin.  Sorry.
> 
> On 6/26/19 1:58 PM, Todor Fassl via Ale wrote:
>> Right, but that is my point. If I run uptime and I see the load on a
>> system is high, I still have to manually figure out if it is cpu bound,
>> memory bound, or disk IO bound, or network IO bound. If you google for
>> tutorials on diagnosing load problems, they all say something like
>> "First run top and look at column 10. Then run iotop and look at column
>> 23. Then run netstat and ..." I don't think I should have to do that in
>> 2019.
>>
>> Surely by now someone has written something to just take a good guess,
>> right? I mean, I could write it myself in perl. Parse the output from
>> top. Then parse the output from iotop. Etcetra. But surely someone  it,
>> has already written that script, right?
>>
>> On 6/26/19 12:35 PM, Lightner, Jeffrey via Ale wrote:
>>> +1 for htop
>>>
>>> It all depends on what you mean by "load".  In UNIX days load averages
>>> in top and other tools were only for the CPU.   In Linux that isn't
>>> the case.  Coincidentally I'd just been in another thread mentioning
>>> that when someone shared this discussion of why that is different in
>>> Linux:
>>>
>>> http://www.brendangregg.com/blog/2017-08-08/linux-load-averages.html
>>>
>>> As far as monitoring goes you could use something like Nagios and the
>>> plugins it provides (or just the plugins and make your own routine to
>>> run the plugin and email you the output).
>>>
>>> -----Original Message-----
>>> From: Ale <ale-bounces at ale.org> On Behalf Of Beddingfield, Allen via Ale
>>> Sent: Wednesday, June 26, 2019 1:26 PM
>>> To: ale at ale.org
>>> Subject: Re: [ale] System Load Summary Script?
>>>
>>> When troubleshooting that type of issue, "htop" is always my
>>> first-glance sanity check.
>>>
>>> Allen B.
>>>
>>> On 6/26/19 12:21 PM, Todor Fassl via Ale wrote:
>>>> Anybody know of a debian/ubuntu package that provides a simple system
>>>> load summary? Maybe you are familiar with mysqltuner. I am looking for
>>>> something like that for system loads. Every time I have a problem with
>>>> a system under high load, I have go google for tutorials on diagnosing
>>>> load problems. Top, iostat, iotop, sar, etc. I'd like something that
>>>> did each of the things these tools do individually and take a best
>>>> guess at what is wrong.
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> https://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo
> 

-- 
Todd