Learn about deploying the same app to App Engine, Cloud Functions, and Cloud Run with no code changes! We’d love to hear what you have questions about in the comments! Subscribe to Google Cloud Tech → goo.gle/GoogleCloudTech Be sure to check out Part 1: Picking the right serverless platform → goo.gle/3PsgUFS
NOTE: the 3rd deployment featured at 6:31 should say "Deploy Node.js app to Cloud Run *via Cloud Buildpacks* " as that's what is used. (It doesn't use Docker [because there's no `Dockerfile` ].)
Hi there, Thank you for this great video :) Just a quick question for my understanding: Only the last sample / #7 has "without Docker via Cloud Buildpacks deployment" in the title. Isn't sample #1, #2, #3, #4, #5 and #6 also using Cloud Buildpacks in the background? This video ua-cam.com/video/suhCr5W_bFc/v-deo.html states that App Engine, Cloud Run and Cloud Functions also build via Cloud Buildpacks. Thank you in advance! #1) 4:45 - Deploy Node.js app to App Engine #2) 5:33 - Deploy Node.js app to Cloud Functions #3) 6:31 - Deploy Node.js app to Cloud Run #4) 7:32 - Deploy Python 2 app to App Engine #5) 8:20 - Deploy Python 3 app to App Engine #6) 9:14 - Deploy Python 3 app to Cloud Functions #7) 10:19 - Deploy Python 2 app to Cloud Run with Docker #8) 11:21 - Deploy Python 3 app to Cloud Run with Docker #9) 12:36 - Deploy Python 3 app to Cloud Run without Docker via Cloud Buildpacks
Congrats for being such a careful viewer! You're close. The buildpacks that build GAE & GCF apps have some proprietary magic, so yes, while they are "buildpacks," it's not "Cloud Buildpacks." Cloud Buildpacks are only used for Cloud Run, meaning the deploys of either the Node.js or Python 3 apps to Cloud Run; the rest use Google-internal buildpacks. Another example of Cloud Buildpacks is using the Functions Framework where Cloud Buildpacks takes your locally-developed Cloud Function and containerizes it for Cloud Run.
Great Knowledge shared in few mints , Cant wait to try these Thanks Wes and Martin!! Question: Can I know which serverless mode would be better for a ML model pickle file picked API and pass the result to another web application or suggest ways to frame Pickle file driven API in GCP
Hi Veena, and thanks for your comment. Your use case sounds fairly simple, straightforward, and consists of just one main piece of functionality (serving pickled ML models via an API), so this sounds like a job for Python 3 Cloud Functions, but you probably guessed that already. App Engine and Cloud Run are better choices for apps. With this video, you can get a prototype running on all 3 to better understand them.
Impressive how simple moving between deployment is in google cloud. That was a good demo. Thank you.
Learn about deploying the same app to App Engine, Cloud Functions, and Cloud Run with no code changes! We’d love to hear what you have questions about in the comments! Subscribe to Google Cloud Tech → goo.gle/GoogleCloudTech
Be sure to check out Part 1: Picking the right serverless platform → goo.gle/3PsgUFS
NOTE: the 3rd deployment featured at 6:31 should say "Deploy Node.js app to Cloud Run *via Cloud Buildpacks* " as that's what is used. (It doesn't use Docker [because there's no `Dockerfile` ].)
Hi there,
Thank you for this great video :)
Just a quick question for my understanding:
Only the last sample / #7 has "without Docker via Cloud Buildpacks deployment" in the title.
Isn't sample #1, #2, #3, #4, #5 and #6 also using Cloud Buildpacks in the background?
This video ua-cam.com/video/suhCr5W_bFc/v-deo.html states that App Engine, Cloud Run and Cloud Functions also build via Cloud Buildpacks.
Thank you in advance!
#1) 4:45 - Deploy Node.js app to App Engine
#2) 5:33 - Deploy Node.js app to Cloud Functions
#3) 6:31 - Deploy Node.js app to Cloud Run
#4) 7:32 - Deploy Python 2 app to App Engine
#5) 8:20 - Deploy Python 3 app to App Engine
#6) 9:14 - Deploy Python 3 app to Cloud Functions
#7) 10:19 - Deploy Python 2 app to Cloud Run with Docker
#8) 11:21 - Deploy Python 3 app to Cloud Run with Docker
#9) 12:36 - Deploy Python 3 app to Cloud Run without Docker via Cloud Buildpacks
Congrats for being such a careful viewer! You're close. The buildpacks that build GAE & GCF apps have some proprietary magic, so yes, while they are "buildpacks," it's not "Cloud Buildpacks." Cloud Buildpacks are only used for Cloud Run, meaning the deploys of either the Node.js or Python 3 apps to Cloud Run; the rest use Google-internal buildpacks. Another example of Cloud Buildpacks is using the Functions Framework where Cloud Buildpacks takes your locally-developed Cloud Function and containerizes it for Cloud Run.
Great Knowledge shared in few mints , Cant wait to try these Thanks Wes and Martin!!
Question: Can I know which serverless mode would be better for a ML model pickle file picked API and pass the result to another web application or suggest ways to frame Pickle file driven API in GCP
Hi Veena, and thanks for your comment. Your use case sounds fairly simple, straightforward, and consists of just one main piece of functionality (serving pickled ML models via an API), so this sounds like a job for Python 3 Cloud Functions, but you probably guessed that already. App Engine and Cloud Run are better choices for apps. With this video, you can get a prototype running on all 3 to better understand them.
Thanks guys great stuff