New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
From my experience almost universally they use nothing, they just wait for complaint tickets to start coming in, and that's how they even discover that something is wrong with a node or the network or etc.
But many vps providers are so fond of suspending services for cpu abuse, how do they come to know of cpu abuse or abuse of other resources?
Depends on node visualization type.
VPSMON works rather nicely for us. Still in beta so has it's issues, but does a great job!
From what I've seen when we acquired hosts, most of them have absolutely no monitoring either for uptime or for resource monitoring.
We use a home grown system which allows us to monitor the server with an agent as well as IPMI (for temperature, reboot)
As we use xenserver as our primary platform, we use citrix xencenter as monitoring solution, and it works great
@drserver - Zenoss has a nice Xenserver Zenpack for this too
Monitor just using proxmox as nodes are all clustered, however I also got a premium pingdom account for uptime. Which reminds me I need to change the settings on pingdom last time one went down I was getting like 20 text messages a minute >.<
@MarkTurner i will check it out. Thank you for suggestion
Isn't VPSMON for OPENVZ only? How do you monitor your KVM nodes?
Correct. KVM is a much smaller portion of our operation at the moment, so that's done more manually.
We utilize WANGUARD for our network monitoring, we also have it setup to monitor for anomalies and nullroute inbound AND outbound attacks via specific packets and/or throughput thresholds that we set -- this all gets passed to the upstream and is nulled at the upstream level, not at the switch level. Most hosts cannot do that though
@drserver - there are some config issues with it, but we have a lot of XenServer 6.2 nodes and they are all monitored with Zenoss and 8 monitoring nodes. Its fantastically good.
Munin, nodeping, zabbix, nagios and eyeballs
Started with Munin and worked great for the first 20-30 nodes, but started failing afterwards with lots of errors till tweaks to draw only the required graphs and worked again for more nodes, but it is very heavy even on SSD.
Nodeping monitors uptime and ssh for everything, zabbix the cloud and other internal services, nagios a few odd chicks here and there and 3 people look at it in turns during the 24 hours.
monitoring? that reminds me of NSA
Zabbix for nodes and simple ping script for virtual machines.
We use Opmananger for monitoring ping, CPU, memory and network traffic externally.
We then use nodewatch one the nodes to catch spammers and other abusers, plus we also manually monitor the nodes pretty heavy to catch things that nodewatch can't.