derek63
commented
10 years ago Start with Turbine stream aggregation per availability zone, will open another issue to aggregate availability zones together
Open jewzaam opened 10 years ago
derek63
commented
10 years ago Start with Turbine stream aggregation per availability zone, will open another issue to aggregate availability zones together
Review netflix turbine github wiki on how to configure turbine.
I assume for configuration PuppetDB can be used, if not drop the first two tasks.
Update this module to:
For reference of java options, see https://github.com/jewzaam/openshift-turbine. The requirement is the turbine configuration needs to know dynamically each node to monitor. PuppetDB is a good way to do this. Also need to maybe support having turbine know about other turbine hosts. So that one stream can aggregate another stream.
Target architecture doesn't have turbine in there, but note that there are multiple availability zones. The goal is to deploy a turbine instance and hystrix dashboard in each AZ:
Therefore the turbine instance in each AZ needs to aggregate streams from each other AZ. My initial thought is have turbine simply pull in the remote AZ's aggregated stream (turbine). If not, then each turbine instance has to aggregate every single metadata and crud node from each other AZ.. a real pain.