a year ago
I migrated my existing app to EU Metal with volume migration enabled. It was originally hosted in US West Oregon. The whole setup migrated smoothly, which was good to see.
However, after the migration, the Redis instance (migrated correctly) started showing increased ping. From the App service in the same EU Metal region, a new PING to Redis now shows ~150-200ms. Before, in US West Oregon, it was few milliseconds (1-10ms).
The App service connects to Redis over public TCP networking. I know using a private URL ensure lowest latency in all cases (experienced in my MongoDB instance), but the Redis driver version I use doesn’t support private adresses and migrate to a newer version requires lot of work. But what I don’t understand is why, in US West Oregon, the same public TCP setup had really small ping, and now in EU Metal, using the same kind of public endpoint and both service and redis in the same region, the latency is much higher.
7 Replies
a year ago
You moved from us to eu? Thats the issue. Try a Metal region in the US
But I supposedly migrated all services, volumes to the EU. I would have expected this to be the same as have created them originally for the EU…
a year ago
are all services now located in EU Metal?
Yes. I mean, the involved ones they are, app service and redis (other like mongo). I have others in other projects or workspaces in US still. But I was not expected to know if this'd be a problem.
a year ago
please have the service connect via the private network instead
Ok, I will try an approach, with MongoDb I have a private network and works well. But in Redis I was limited to a version, but got another idea. Lets see. Thank you