What mechanisms keeps the participant to domain connection alive or automatically reconnecting?
When the two servers are on remote computers the network connection might be severed, is the ping described here the recommended way to make sure my participant is able to process commands that involve a given domain?
Thank you
Generally, the participant will automatically keep on reconnecting until the domain connection comes back. If you type participant.health.status you should the domain connection as unhealthy if there is an outage.
The /health HTTP end-point is a bit too dumb at the moment and the only way to expose an issue with the domain on the /health endpoint is the ping check you are citing.
We are working on improving that /health endpoint to give much better signals and conditions, telling the admin what exactly is broken. But that’s not available yet.
@Ratko_Veprek Thanks for that! @Mate_Varga also pointed me to this keep-alive documentation , so I should configure that for all my participant and domain nodes too?
You don’t need to configure it. It’s enabled by default and I would recommend to not change the default settings unless you experience problems that can somehow be linked to it (I wouldn’t know what). The documentation is here simply for completeness.
If you change the settings, please consult the GRPC manual beforehand so that you understand exactly what you do. If you configure it wrong, GRPC servers will shut your client connections …