In the context of distributed services it would be nice to simulate different latencies between groups of containers. For example, instead of providing the filtering IP and list of containers, the user can provide a graph file where different groups are defined
pumba netem delay --graph graphfile
and the graph file can be something like this
groups:
group_1:
container_1_1
container_1_2
group_2:
container_2_1
container_2_2
ip_2_3 # you can support both IPs and container names
group_3:
container_3_1
filters:
filter_1:
src_group: group_1
dst_group: group_2
delay: 150
filter_2:
src_group: group_1
dst_group: group_3
delay: 250
In the context of distributed services it would be nice to simulate different latencies between groups of containers. For example, instead of providing the filtering IP and list of containers, the user can provide a graph file where different groups are defined
and the graph file can be something like this