I am in the process of building few utilities for a in-house DC setup consisting of few servers of varoious types ( in service, brand and os's). The total number of servers is 25, and all of them are for internal use only, so nothing to with customer or billing management.
Some things that I have already built are:
Service monitoring, including response time monitoring for few services like http, DB(Oracle and MS-SQL) connection, Email Server etc.
Physical Server availability through Ping response times.
Server monitoring for few critical items like Processor Memory and bandwidth Usage.
Alerts to admin groups in case of problem
What else can be added to this small selection? Please keep in mind that it is a small in-house dc facility, and the software is for the benefit of admins only. If possible, please stick to software management utilities only, since I cannot do anything about the hardware or the infrastructure setup of the dc.
Thanks guys, but I am not sure if all of them are pure software parts. Actually I am only concerned with the management part through software, not in the running of the dc. I am sorry for this since I was not clear in my first post. I will edit it now.
I already have bandwidth monitoring. For the power backup, I will have to see whether APC has any interface for third party developers or not to measure remaining backup battery time. For the IP management, although this may not be that useful for a small setup, I will add it at a later stage.
I am sorry, i may misunderstood. You want the infomation like what are the things need to be involved when designing software to manage datacenter?. Please let me know, let see what are the information I can find it for you.
It is really intresting. I will make a list and let you know soon. Also do you like to calclualte BW?, Do you take care billing issues too or onaly managing servers.
assigning multiple IP addresses to server
What kind of BW controler you would like to intergrate? MRTG
email to client if BW exceeded
IP Management
Device Management ( Routers, Swithces, server Hardware, etc...)
Server Configuration
Server Location
Reboot Management using KVM switches
Check single bandwidth chart for a port on your switch
Also you can have
Support Management & Billing Management.
For Bandwidth management, we are already using MRTG. Planning to shift to SNMP polls within the next few days.
Haven't touched IP management, but that's an intresting point to include. Actual assigning of IP's to the server would require a client to be installed on the server, which I wan't to avoid.
Device Management is another intersting thing. Some of the info can be had through SNMP. Will check it as a first priority.
Changing server configuration again requires having a client, although viewing might just be possible through SNMP.
The DC where I am working doesn't have a KVM switch, otherwise I would have surely included Reboot Management in my modules.
I will have to take a look on how to monitor a switch port for bandwidth. I suppose it should be again through SNMP.
As for support and management, I am not sure that is immediate priority, as I am in a private DC right now.
Thanks for all the info. It will certainly help me in adding many more things to my app.