Difference between revisions of "Ceilometer/MonitoringPhysicalDevices"
m (Text replace - "__NOTOC__" to "") |
|||
(4 intermediate revisions by one other user not shown) | |||
Line 1: | Line 1: | ||
− | + | ||
= Monitoring Physical Devices = | = Monitoring Physical Devices = | ||
+ | == Goals == | ||
+ | It should be possible to monitor physical devices in the [[OpenStack]] environment. The monitored devices are: | ||
− | |||
− | |||
− | |||
− | |||
− | |||
* the physical servers on which Glance, Cinder, Quantum, Swift, Nova compute node and Nova controller runs | * the physical servers on which Glance, Cinder, Quantum, Swift, Nova compute node and Nova controller runs | ||
* the network devices used in the [[OpenStack]] environment (switches, firewalls ...) | * the network devices used in the [[OpenStack]] environment (switches, firewalls ...) | ||
+ | == High Level Description == | ||
There is a new Ceilometer agent needed to achieve this. This agent should run on every physical [[OpenStack]] server. To collect data there should be a way to define and support plugins to collect the data(SNMP, Proc, IPMI ...). For example, to monitor network devices we can simply add and configure a SNMP extension on any physical [[OpenStack]] server. | There is a new Ceilometer agent needed to achieve this. This agent should run on every physical [[OpenStack]] server. To collect data there should be a way to define and support plugins to collect the data(SNMP, Proc, IPMI ...). For example, to monitor network devices we can simply add and configure a SNMP extension on any physical [[OpenStack]] server. | ||
− | == | + | Picture of architecture: [[attachment:architecture.jpg]] |
+ | |||
+ | Class diagram: [[attachment:Class Diagram]] | ||
+ | |||
+ | == User stories == | ||
+ | Questions that must be answered: | ||
+ | |||
+ | * Network | ||
+ | ** Is enough network bandwith aviable? | ||
+ | ** What is the trend of network usage? | ||
+ | ** How much bandwith does a customer use? | ||
+ | ** How much internet bandwith does a customer use? | ||
+ | * Compute | ||
+ | ** Is enough computing power aviable? | ||
+ | ** How much computing power does a customer use? | ||
+ | ** Is enough RAM aviable? | ||
+ | ** How much local storage does a customer use? | ||
+ | * Storage | ||
+ | ** Is enough storage aviable? | ||
+ | ** How much storage does a customer use? | ||
+ | * Miscellaneous | ||
+ | ** How many new customers can be served? | ||
+ | ** What are the temperature of the components? |
Latest revision as of 23:30, 17 February 2013
Monitoring Physical Devices
Goals
It should be possible to monitor physical devices in the OpenStack environment. The monitored devices are:
- the physical servers on which Glance, Cinder, Quantum, Swift, Nova compute node and Nova controller runs
- the network devices used in the OpenStack environment (switches, firewalls ...)
High Level Description
There is a new Ceilometer agent needed to achieve this. This agent should run on every physical OpenStack server. To collect data there should be a way to define and support plugins to collect the data(SNMP, Proc, IPMI ...). For example, to monitor network devices we can simply add and configure a SNMP extension on any physical OpenStack server.
Picture of architecture: attachment:architecture.jpg
Class diagram: attachment:Class Diagram
User stories
Questions that must be answered:
- Network
- Is enough network bandwith aviable?
- What is the trend of network usage?
- How much bandwith does a customer use?
- How much internet bandwith does a customer use?
- Compute
- Is enough computing power aviable?
- How much computing power does a customer use?
- Is enough RAM aviable?
- How much local storage does a customer use?
- Storage
- Is enough storage aviable?
- How much storage does a customer use?
- Miscellaneous
- How many new customers can be served?
- What are the temperature of the components?