Server Monitoring & Notification: Page 2 - Enhance Control Panel

Server Monitoring & Notification

xyzulu

jeff What do you mean by integration? You just install netdata and that's it.. for a free service I don't think netdata is going to easily allow their dashboard to be shown in enhance.. (yes, there are paid options).

kelsallp

Has anyone set up the LiteSpeed plugin? Where do the outputs appear?

xyzulu

What are you hoping to see/monitor with the Litespeed plugin?

kelsallp

Just experimenting on what it can give after reading this

kelsallp

https://www.netdata.cloud/litespeed-monitoring/

tecreasey

I had to remove netdata when I last tried it because is caused resource spikes and almost rendered one of my nodes inaccessible.

Upon uninstalling, the resources returned to normal.

I would love to use something like netdata to get more details about the performance of my nodes.

kelsallp

tecreasey Interesting - I have noticed a significant jump in resource usage after installation of the Linux monitoring module.

Hennie

just installed netdata. oh man, so much data in few minutes. 🙂 seems to be a great monitoring solution.

necrom

twest We had an issue recently where load would jump to the 100's twice a day, turned out to be an issue with all versions of ESET10 and it was fixed with a downgrade to latest ESET9.

twest

necrom is that a security app for servers? Haven't heard of it, appreciate the tip though! I might just need to wait for the next couple Enhance updates that includes more logs and systems resources charts, people got me squirrely about using netdata now lol... I did read it requires a good amount of resources, which makes sense considering the level of detailed resources monitoring and by-the-second. I just can't risk a server locking up on me, so prolly gonna wait and see what Enhance comes out with.

Rich

@twest I saw it was using about 5- 12% cpu, memory is very small. >200mb io will be higher as it stores so much but again nothing excessive standing out.

If you're in need of deep analysis of time series recorded data it's the tool, else it's a bit bloated in it's detail it captures. Or strip out all the bloat you don't want capturing and try and make it lighter?

Also, what about Zabbix/Nagios or Prometheus as another alternative? Muin was always welcomed on my cPanel servers, but Netdata got installed when Muin wasn't enough to drill into some tricker issues.

twest

Rich yeah I am considering a Zabbix install. I see they have mobile apps available now, so monitoring and notifications are extra convenient.

My main issue at the moment has to do with backups causing load spikes, so I'm doing some tests on those to see if I just have them set too aggressively.

Long-term I'll definitely be wanting something a little more robust for monitoring and notifications for sure though, so Zabbix will probably be my next step soon.

8Dweb

twest I saw your other thread on this. Question, is the backup service running on a single install of enhance, or a VM running on the same machine? Are the backups going to local disk/array or to alternative disk/storage medium?

twest

8Dweb the backup service is on a dedicated server, has tons of CPU from dual xeons, 64GB ram, has NVMe storage, has 10Gbps port, no bottleneck to be found for the backups server I'm sure. Load monitor on the backups server is minimal too, no spikes there.

Rich

Just to share an observation. One thing I noticed with the backups is they usually never take up any resources, I don't even see any bumps in the graphs for most small sites. But I do have a >100GB multi-press + >100GB NextCloud in a different estate that it's backups make the backup sever load jump high for about an hour, for each backup. It is having to process >2 million files, but it's very obvious... An underpowered VPS on a bad network made that 4x times longer than it is now. Now it's a VPS with dedicated attached SATA disk. So the bottle neck could be the IO of the sata drive.

Load on the main sever where the website sits, isn't even detectable, it's only the backup server side. So IO processing? The BTRFS Snapshot diff's probably.

twest

Rich I'm not using sata on any servers, and they're all dedicated servers, only using enterprise NVMe, which has insane IO, million+ IOPS/s, 30+ Gbps transfer.

That's what makes it so odd to get load issues, there's no obvious bottleneck. I recently reduced the number of concurrent backups from 10 to 5 and reduced from every 3 hours to 4 hours, and that seemed to help.

I also lowered the default Linux swapiness from 60 to 30 to see how that fairs.

ivansalloum

twest From which provider? Hetzner?

Rich

Fingers crossed for you bud.

« Previous Page