Sysadmin monitoring and alert tools?

Let me join the club

---

 

Here is mine for the past 24 hours:


Linode 4096 plan.

Average ~2000 users daily according to my Google Analytics "Audience Overview"

17K members,

530K posts.

I don't know if this is good or bad honestly. This is the only site (XF 1.5.x) I have on this box right now.

PHP 7.1.12
Server version: 10.0.33-MariaDB MariaDB Server

 

Wow, really interesting to see how everyone's graphs compare!

I am drooling at your low PHP times, @eva2000!

I bet it's more than just traffic, though... my PHP-FPM config is probably leaving a lot of performance on the table. Another issue is probably WordPress — I'm using a good caching plugin, but it's still a real slog when it comes to performance.

For XenForo, though, I wonder what's going on... when it's running at its best, my own forum usually feels faster than the Centminmod Community forums (my own forum actually feels faster than any other XenForo forum I've used).

As a sanity check, I just compared with WebPageTest and it actually seems to support this. Ignoring the full page render time (because of all my ads), the time-to-first-byte on my forum home page is half that of Centminmod's (218ms vs 533ms), and my page's HTML was received in 257ms compared to CC's 581ms.





This is surprising considering the vast difference that NewRelic/AppOptics shows in PHP application time.

The only explanation I could think of is network RTT and proximity? My dedicated server is with ReliableSite in their NYC datacenter (I'm not using CloudFlare or a CDN). Not sure which Centminmod Community server myself or the Dulles VA WebPageTest server would be connecting to?

 

Sorry for a bad picture.
This is from my Nextcloud vps which is only for a family. One user uploading 150 pictures with mobile app and application's graph hits the roof. Should I be worried? So far everything looks ok except that graph.
Redis graph's lower line is 5 and upper 10.
Picture's upload took about 1 minute but application graph was really high about 40 minutes...

 

Even so long after upload?

 

My APM Stats
https://community.centminmod.com/threads/appoptics-apm-xenforo-1-5-response-time-breakdown.14279/

Feel free to share yours .

 

Hello,

at the moment I am looking around for a good monitoring tool, mainly for centOS at the first place, but also for some debian-based servers.

I was thinking about ELK Stack or Graylog, but I dont know much about it and I would like to get a centralized monitoring system which is also secure. I have the feeling adding ELK / Graylog adds another layer of a attack surface to my servers.

So what do you guys use for monitoring multiple servers?

Thanks!

 

Aww damn.. Could you merge this thread here @eva2000 ?

Btw, any opinions on ELK / Graylog? Especially when having multiple servers and setting it up? Seems like pretty messy compared to amplify for example (which btw looks pretty nice, just configured it on one of my servers)

 

Thanks!

Yeah, I dont know.. seems overwhelming to me at the moment. I monitored by SSH into the servers and check the logs, processes, etc. So I was looking for something.. 'better'?!

But all the tools you listed are doing a great job so far and they arent even that expensive, Hetrix or Nixstats for example. Easy to setup and no reading around forums, stack overflow and reddit about how to setup ELK/Graylog and then forward the logs to another server, which makes things PRETTY complicated for me at least.

 

How good is Hetrix? Pros/cons?

 

Oh, I almost forgot about this thread! Surprised there isn't more discussion around NewRelic. Since they updated their pricing structure it's very affordable (and may even be free depending on your site usage).

I am paying around $4/month for APM and server monitoring and the tools are fantastic, far better than any other service I have seen out there (most of which cost a LOT more). It was also very easy to setup with Centminmod.

https://newrelic.com