How do you monitor your vps uptime?

[researchlabs]

I tried uptimerobot and statuscake, the first one is still telling me that a site of my own is online while the domain is expired and totally down since one month ago, the second one think that a server that I monitor thru ip is down since a week ago when it's actually up and working fine.

Do you know any reliable method?

 

[kcaj]

I use UptimeRobot and find it reliable. Other than that I use my own Observium installation but wouldn't really consider it reliable.

 

[WebSearchingPro]

Well, statuscake does ping tests I believe it also caches DNS resolutions which would be why this happened in your case. Their paid option has a live content testing feature that would have resulted in a proper alert.

Now for the false down, its likely a reactive firewall (CSF/LFD on cPanel) ended up blocking the IP after many requests/pings in a short period of time, I've had this happen in the past for http tests. For ping tests if you are using cPanel hosting, your provider may have opted to block ICMP.

 

[researchlabs]

I don't use cPanel but Vesta panel, but this kinda problems are not within my reach now, I have to do experience.

I'll take a look to observium.

 

[jvkz]

We use monitor.us

Best thing about it is you can monitor ram, cpu, bandwidth usage in real time from any browser and last 24 hour monitoring plan is free too.

 

[MannDude]

NodePing and Observium

 

[tonyg]

I use xymon for uptime and server health in addition to a custom bash script for just uptime monitoring and notification.

 

[VMBox]

Not so much for VPS, but our whole infrastructure we use Nodeping and have Pingdom alerts as a backup.

I much prefer Nodeping over Pingdom though. 

 

[Nett]

Nodequery

 

[gonggo]

port-monitor.com - 1 min interval

sentineltower.com (invite only) - 1 min interval

nodequery - 3 min interval

 

[HalfEatenPie]

observium, cacti, and munin.  

I'm currently thinking about moving over to opennms but I haven't taken the time yet to fully investigate it.  

 

[GreenHostBox]

When I had a VPS, I used to use StatusCake and UptimeRobot. They're both very reliable in terms of their free plan in my opinion.

 

[DomainBop]

Internal:

1. Observium

2. N2 Monitoring

3. Zabbix

4. Webmin's built-in status monitoring (useful for auto restarting crashed services in the middle of the night)

5. Custom bash script that monitor's Apache/server load and will auto restart Apache if load is over xx.xx for x minutes (useful for when a server is attacked by a bunch of random evil attack bots)

6. Mojeda/ServerStatus script installed on various VPS's on every continent (useful for a quick visual check)

External:

1. Monitor.us (free version <--30 minute checks mean it's already old news when they send out alerts )

2. StatusCake (paid version)

Big Brother:

1. OVH/Kimsufi's RTM monitoring servers there

2. CloudVPS N2 monitoring things there

 

[kcaj]

1. OVH/Kimsufi's RTM monitoring servers there

Apparently Kimsufi don't use the OVH RTM tool, they only ping the server to check it's online. I can't see it running on my Kimsufi.

 

[DomainBop]

Apparently Kimsufi don't use the OVH RTM tool, they only ping the server to check it's online. I can't see it running on my Kimsufi.

Kimsufi uses RTM.  It's installed at /usr/local/rtm (on Debian) and runs a cron job every 60 seconds.  The run time is only a second when the cron runs so you probably won't notice it in a list of running processes.

If you check your syslog you should see a bunch of entries that look like this:

May 26 12:05:01 marseille /USR/SBIN/CRON[15790]: 
(root) CMD (/usr/local/rtm/bin/rtm 57 > /dev/null 2> /dev/null)

 

[kcaj]

Kimsufi uses RTM.  It's installed at /usr/local/rtm (on Debian) and runs a cron job every 60 seconds.  The run time is only a second when the cron runs so you probably won't notice it in a list of running processes.

If you check your syslog you should see a bunch of entries that look like this:


May 26 12:05:01 marseille /USR/SBIN/CRON[15790]:
(root) CMD (/usr/local/rtm/bin/rtm 57 > /dev/null 2> /dev/null)

Ah yes, will take a look at the scripts in there now.

I had asked on the forum yesterday but was told only ping is checked.

 

[WSWD]

Oh...you don't even want to get me started on Statuscake.  Complete and utter load of crap.  Here's a nice lengthy thread I started over on WHT:  http://www.webhostingtalk.com/showthread.php?t=1355681

I have been with Nodeping for a few months now, and absolutely love them.  I started getting a few false positives about a week ago, but Shawn (the owner?) worked with me, and I believe we have things resolved now.  It was just a matter of increasing the ping response time to 15 seconds.  Apparently one server was not responding to certain Nodeping test nodes in a timely manner.

Aside from that, I have never had a false positive with them, and have never had them miss an outage.  Been very happy since I switched.

 

[Mid]

hyperspin.com

not used them, but my provider used (and still uses) them when I was on shared hosting

pricing seems to be moderate on 3 min+ intervals

 

[DomainBop]

Oh...you don't even want to get me started on Statuscake.  Complete and utter load of crap.  Here's a nice lengthy thread I started over on WHT:  http://www.webhostingtalk.com/showthread.php?t=1355681

You must be mistaken.  StatusCake is 100% accurate and never has any problems.

Fatal error: Uncaught exception 'MongoCursorTimeoutException' with message '176.227.201.226:27017: Read timed out after reading 0 bytes, waited for 30.000000 seconds' in /home/Sites/StatusCake/App/Workfloor/Get.Hourly.Performance.php:12 Stack trace: #0 /home/Sites/StatusCake/App/Workfloor/Get.Hourly.Performance.php(12): MongoCursor->rewind() #1 /home/Sites/StatusCake/App/Workfloor/ListSites.php(74): GetHourlyPerformance('157897') #2 /home/Sites/StatusCake/App/Status.php(142): ListSites() #3 {main} thrown in /home/Sites/StatusCake/App/Workfloor/Get.Hourly.Performance.php on line 12

 

[markjcc]

I like to use statuscake as it provides a public access page for any new potential clients to see