Deployment yaml global - params

Community Support
,
1

Hi,

My Temporal-server running on k8s and running each service as individual pod.
image

global:
membership:
maxJoinDuration: 300s
broadcastAddress: “##TEMPORAL_POD_IP##”
pprof:
port: 7936
metrics:
prometheus:
timerType: “histogram”
listenAddress: “##TEMPORAL_POD_IP##:8000”
clusterMetadata:
enableGlobalNamespace: true
failoverVersionIncrement: 10
masterClusterName: “active”
currentClusterName: “active”
clusterInformation:
active:
enabled: true
initialFailoverVersion: 1
rpcName: “frontend”
rpcAddress: “##TEMPORAL_POD_IP##:7233”

Currently I have configured the TEMPORAL_POD_IP ie., current pod ip address (the screenshot you see is for history pod address). If I see the deployment.yml from matching pod you can see matching ip address the same for others.

image
image

I would like to know the expected exact ip addresses (dynamic) for the broadcastAddress, listenAddress & rpcAddress ?

2

broadcastAddress is the address used by ringpop to communicate membership info. In this case, it should be your pod ip address, you did it right.
listenAddress is the address where temporal server opens up endpoint for prometheus scraper. Your config is correct. But you could also use form of “0.0.0.0:8000”. You will need to config your prometheus to scrap from port 8000.
rpcAddress is the address used by server’s worker role to connect to current cluster. Usually, you use the same dns name that you would want your client to connect to. This config key is deprecated and is replaced by

publicClient:
  hostPort: server_frontend_dns:7233

Server code might still use both rpcAddress and publicClient.hostPort. We will clean it up soon so it only needs publicClient.hostPort.