Speed up your Cloud Run service by hosting it closer to your users

When it comes to response times on your app, every millisecond counts. In this episode of Serverless Expeditions Extended, Martin and Ahmet look at why Cloud Run may be slower for users in different regions. They also show how to use the Cloud HTTPS Load Balancer and Global Anycast IPs, so wherever the user connects from, they are routed to the "nearest" Google data center.

Timestamps:
0:00​ - Intro
1:08 - Why a Cloud Run service may be slower for some users
1:49 - How to bring your Cloud Run service closer to users
3:02 - Testing the service after bringing it closer to users
4:20 - Testing the service from another region
4:50 - Terraform and wrap-up

#ServerlessExpeditions​ #ServerlessExpeditionsExtended #CloudRun

Product: Cloud Run; fullname: Martin Omander, Ahmet Alp Balkan;

[burakk]

Seeing a Turkish engineer in Google, explaining a cutting edge solution about an important point made my day. Thank you Ahmet.

[rainedJMTi]

oh okay, we can deploy cloud run service with multiple regions under the same service name, this should be highlighted. I was trying to figure out how do I handle the load balancer configuration to point to multiple backend services.

[ricardocuenca]

What about services with a db? Firestore/datastore can't be replicated (please let us do that!), Cloud SQL doesn't do that out of the box (I think). In fact only Cloud Spanner has multiregion replication, but it's so expensive and has stronger consistency than many apps might need. Is there a not-too-complicated, ideally managed way to distribute db-backed apps globally?

[rao]

good services providing by cloud service platform!

[orinda.harrison]

Is it possible to use memcache with cloud run?