Agree that we should have at least 1 implemented monitoring plugin. Before we jump into SNMP etc, I'd like to see the public interfaces of the monitoring and AS services. In the figure I had sketched out, I'd like to see the specifications of all the APIs.
We can then enumerate the autoscaling workflows and see if the APIs are sufficient. Some of the flows are: 1. Create a launch configuration 2. Create AS group using the launch config in #1 3. Scheduled scaling 4. Dynamic (load-based) scale up and down On 12/3/13 2:54 AM, "Daan Hoogland" <daan.hoogl...@gmail.com> wrote: >On Tue, Dec 3, 2013 at 10:03 AM, Nguyen Anh Tu <t...@apache.org> wrote: >> + what SNMP software we use? I found Net-SNMP but not sure it's okie. >>Need >> to try and review it. >There are a lot of oopen source and commercial once out there. We need >an open source one. Given that we are probably going to have to do >some work on the library our selves, I think license compatability >with apache deserves the greatest consideration. > >> + what metric we decide to monitor for AS action? >I think we should design a solution that reads from a configuration >file what mibs/oids to query from what device. So we don't need to >make this decision in this (design) phase
