Bandwidth reduction

Hi,
We are using consul in very big cluster deployments on AWS.
Each deployment is using a configuration of 3 master machines each on a different AZ so we can have fault tolerance.

Lately we saw that we are paying very big amount of money because of traffic between AZs.

We tagged our resources and we can say for sure that the problem is traffic from/to managers.
How can we minimize this traffic? which parameters are relevant for this?

Thank you for your time,

Ofer

Hi @OferE,

Without knowing much about your environment it is hard to offer recommendations. Would you mind providing a bit more info about your environment such as how many servers are deployed within a single Consul data center, and whether Consul being used for service discovery, KV store, or both?

Assuming Consul is being used for service discovery, you may want to consider allowing Consul to send stale responses for DNS (https://www.consul.io/docs/agent/options.html#max_stale) or HTTP requests (https://www.consul.io/docs/agent/options.html#discovery_max_stale) – if these are not already enabled. Enabling these options would have an affect on bandwidth utilization.