Background
It is very important that with the volume of network data being collected, and the innovative mechanisms to detect an 'event', that scalable, flexible and secure mechanisms of integrating alerting mechanism be investigated.
Alerting systems should be clever and configureable enough to handle any kind of input and be able to send alerts through various forms of communication: eg email, paging systems, web pages, SNMP traps, etc. Also as important, a archiving/historical database of alerts should investigated and it's design tested.
Specifics
Different types of reports will be needed for various different end users of such alerts. a clearly defined api should be created that is flexible to configuration and reliable. Messages generated should be concise enough that the person receiving the message is able to determine what is wrong, yet verbose enough to understand the specifics of the problem. In conjunction with the history archive, mechanisms to quickly retrieve the full information for the alert should be also available.
Existing open source alerting systems may already be in existance. they should be understood and reviewed for the suitability of this project.
Goals
- understand the use cases of network alerting
- develop a consistent API framework for alerting. could potentially be a daemon like syslog.
- implement said API, with considerations for the different types of alerting mechanisms possible: SNMP traps, email are two important mechanisms.
- Work with Event Diagnostics projects to determine the best way of integrating the reporting with alerting.
Contact
Yee.