Google Cloud Technology Nuggets — Dec 16–31, 2022 Edition

Romin Irani
Google Cloud - Community
5 min readDec 30, 2022


Welcome to the December 16–31, 2022 edition of Google Cloud Technology Nuggets.

Top stories blogs

Towards the end of the year, blog posts that round up the top stories of the year or even predictions can be interesting reads as you set yourself for work in the new year. In no order of preference, here are some blog posts in this category:

  • This entire newsletter is derived from the official Google Cloud blog, so it makes sense to highlight the post that features Top 22 news stories of 2022, as per the readers.
  • Check out the Top 10 Transformation stories for the year 2022. Some interesting stories in this list look at how Golden State Warriors have harnessed data, AI predictions for the next decade, lessons learnt from Google’s Ares 120 Incubator and more.
  • At Google Cloud NEXT ’22, 10 of our experts shared predictions of what to expect in IT till 2025. A blog post in early December shared those insights.
  • The adoption of low-code and no-code tools is here to stay. Check out this prediction that says that by 2025, professionals that do not identify themselves as professional developers today will build half of all business applications.


The CISO Perspectives for December is an interesting post, which in addition to its standard roundup of security news for the month, provides top security lessons that have been learnt in 2022. Check it out.

In case you prefer Podcasts, it’s worth mentioning that there is a Cloud Security Podcast that publishes monthly episodes.


Our Startup Technical Guides are focused on startups taking their journey though the start, build and grow cycles of their organization. Check out this blog post that highlights key videos and tutorials that help you in the build phase. The videos include deep dives in security, BigQuery, Data Processing and more.

Storage, Databases and Data Analytics

Do you need better insights in your BigQuery analytics query? Facing possible performance bottlenecks with them? You can now tap into query execution graph that can provide you with those insights.

Check out the blog post for more details.

Memorystore is a fully managed implementation of the open source in-memory databases Redis and Memcached. Check out this blog post that discusses best practices for using Memorystore. These practices include optimizing memory settings, configuring maintenance windows, high availability setup, configuring alerts/metrics and more.

Machine Learning

Document AI has announced 3 OCR features in Public preview:

  • You can now access a document for its Intelligent Document Quality (IDQ) score, which is measured via page level quality metrics across various dimensions. These include Blurriness, Darkness, etc.
  • Digital PDF support in Document AI OCR
  • OCR versioning, which allows a user to pin a specific OCR Model to lock its behavior and not see any side effects due to upgrades in a model.

Check out the blog post for more details.

Speech AI at Google saw a significant number of announcements. Here is a blog post that recaps some key announcements in Speech AI across 2022.

If you are looking at training models with large amounts of data, the general recommended best practice is to do distributed training and reading from Cloud Storage. The challenge is that if there are multiple files, then they are data throughout and network overhead challenges. Check out this blog post that demonstrates various distributed training paradigms to perform efficient training.


When systems fail, your users do not care about your infrastructure problems? They are focused on the fact that something that they were expecting to work, is not working? How do we shift the focus then from what we have been used to in typical monitoring metrics to instead the ones that the users care about. This blog post makes the case for that.

Developers and Practitioners

Cloud Functions is a popular FaaS option available on Google Cloud. Since Cloud Functions represents an abstraction that is typically run in the Cloud in an environment that is fully managed by the Cloud Provider. This makes it extra difficult to develop your functions locally, simulate a cloud like environment locally for both development and testing.

This blog post gives you fantastic details on how you can go about and set up your local environment for development and testing of functions code.

Cloud Build is Google Cloud’s fully managed CI/CD pipeline tool. You can use Cloud Build to pretty much automate a variety of tasks via provided and community builders. The typical use case is to kick off the build process to create the container image, push the image to a repository and then potentially trigger the deployment process too. In this definitive guide to Cloud Build, check out the step by step goal oriented approach that takes you through the mechanics of understanding Cloud Build in detail by triggering the build process manually and then reaching a state of automation by using specific triggers to kick off the build process. While going through this, you will also get familiar with various services like Cloud Build, Artifact Registry, Cloud Run and Cloud Pub/Sub.

Stay in Touch

Have questions, comments, or other feedback on this newsletter? Please send Feedback.

Looking to keep a tab on new Google Cloud product announcements? We have a handy page that you should bookmark → What’s new with Google Cloud.