Since a common pattern is emerging, it would be useful for me and others if a fly github repo for each was setup to help everyone with this eventually… Everyone is short on time i know… But it just seems like such a standard 2 things so many people need. Metrics and Workflows.
This is fairly simple actually (at least with a non-clustered, singleton temporal deployment with all services in one app, but likely also when deploying the individual services separately).
Note: This is obviously the bare minimum… Without metrics, no externally exposed port for the UI, etc. (probably not desirable but depends on your use-case), no health checks, etc. … But it works as a starting point
Hi, anyone have success w/ the latest temporal images? I’m able to run the temporal server + pg db but I’m having trouble getting clients to connect to it.
I have POSTGRES_* env variables set in the Dockerfile and include ENV BIND_ON_IP=0.0.0.0 EXPOSE 7233
I can ssh into the temporal app and run tctl cluster health => SERVING
When I try to connect to the flyapp, the connection times out, eg: grpcurl -plaintext myapp.fly.dev:7233 list
The temporal client should set the connection option: tls: true. This is just the encryption from the client to the fly server, which will downgrade the tls connection for its internal connections.
But… you probably don’t want to expose your Temporal Server to public access… so don’t include the configuration above and have your web server communicate w/ the temporal server internally via your-temporal-server.internal:7233 without the tls prop.