Posts from August 2020

New Case Studies About Google’s Use of Go

Thursday, August 27, 2020

Go started in September 2007 when Robert Griesemer, Ken Thompson, and I began discussing a new language to address the engineering challenges we and our colleagues at Google were facing in our daily work. The software we were writing was typically a networked server—a single program interacting with hundreds of other servers—and over its lifetime thousands of programmers might be involved in writing and maintaining it. But the existing languages we were using didn't seem to offer the right tools to solve the problems we faced in this complex environment.

So, we sat down one afternoon and started talking about a different approach.

When we first released Go to the public in November 2009, we didn’t know if the language would be widely adopted or if it might influence future languages. Looking back from 2020, Go has succeeded in both ways: it is widely used both inside and outside Google, and its approaches to network concurrency and software engineering have had a noticeable effect on other languages and their tools.

Go has turned out to have a much broader reach than we had ever expected. Its growth in the industry has been phenomenal, and it has powered many projects at Google.
Credit to Renee French for the gopher illustration.

The earliest production uses of Go inside Google appeared in 2011, the year we launched Go on App Engine and started serving YouTube database traffic with Vitess. At the time, Vitess’s authors told us that Go was exactly the combination of easy network programming, efficient execution, and speedy development that they needed, and that if not for Go, they likely wouldn’t have been able to build the system at all.

The next year, Go replaced Sawzall for Google’s search quality analysis. And of course, Go also powered Google’s development and launch of Kubernetes in 2014.

In the past year, we’ve posted sixteen case studies from end users around the world talking about how they use Go to build fast, reliable, and efficient software at scale. Today, we are adding three new case studies from teams inside Google:
  • Core Data Solutions: Google’s Core Data team replaced a monolithic indexing pipeline written in C++ with a more flexible system of microservices, the majority of them written in Go, that help support Google Search.
  • Google Chrome: Mobile users of Google Chrome in lite mode rely on the Chrome Optimization Guide server to deliver hints for optimizing page loads of well-known sites in their geographic area. That server, written in Go, helps deliver faster page loads and lowered data usage to millions of users daily.
  • Firebase: Google Cloud customers turn to Firebase as their mobile and web hosting platform of choice. After joining Google, the team completely migrated its backend servers from Node.js to Go, for the easy concurrency and efficient execution.
We hope these stories provide the Go developer community with deeper insight into the reasons why teams at Google choose Go, what they use Go for, and the different paths teams took to those decisions.

If you’d like to share your own story about how your team or organization uses Go, please contact us.

By Rob Pike, Distinguished Engineer

Recapping major improvements in Go 1.15 and bringing the Go community together

The Latest Version of Go is Released

In August, the Go team released Go 1.15, marking another milestone of continuous improvements to the language. As always, many of the updates were supported by our community of contributors in collaboration with the engineering team here at Google.

Following our earlier release in February, the latest Go build brings a slew of performance improvements. We’ve made significant changes behind the scenes to the compiler, reducing binary sizes by about 5%, and improving building Go applications to be around 20% faster and requiring 30% less memory on average.

Go 1.15 also includes several updates to the core library, a few security improvements, and much more–you can dive into the full release notes here. We’re really excited to see how developers like you, ranging from those working on indie projects all the way to enterprise devs, will incorporate these updates into your projects.

A few users have been working with the release candidates ahead of the latest build and were kind enough to share their experience.

Wayne Ashley Berry, a Senior Engineer at Over, shared that “...seeing significant performance improvements in the new releases is incredible!” and, speaking of compiler improvements, showed “one of [their] services compiling ~1.3x faster” after upgrading to Go 1.15.

This mirrored our experience within Google, compiling larger Go applications like Kubernetes which experienced 30% memory reductions and 20% faster builds.

These are just a couple examples of how some users have already seen the benefits of Go 1.15. We’re looking forward to what the rest of the gopher community will do with it!

A Better Experience For Go Developers

Over the last few months we’ve also been hard at work improving a few things in the Go ecosystem. In July, the VS Code extension for Go officially joined the Go project and more recently, we rolled out a few updates for our online resources.

We brought a few important changes to, a central source of information for Go packages and modules. With these changes came functional improvements to make the browsing experience better and minor tweaks across the site (including a cute new gopher). We also made some changes to—our hub for Go developers—making it easier to navigate the site and find examples of Go’s use in the enterprise.
The new home page on Credit to Renee French for the gopher illustration.
We’ll be bringing even more improvements to the Go ecosystem in the coming months, so stay tuned!

Our Commitment to Open Source and Google Open Source Live

Most of these changes wouldn’t be possible without contribution from our open source community through submitting CLs to our release process, organizing community meetups, and engaging in discussions about future changes (like generics).

Being part of the open source community is something that the Go team embraces, and Google as a whole works to support every year. It’s through this community that we’re able to iterate on our work with a constant feedback loop and bring new gophers into the Go ecosystem. We’re lucky to have the support of passionate Go advocates, and even get to celebrate the occasional community gopher design!

That being said, this has been a challenging year to gather in person for meetups or larger conferences. However, the gopher community has been incredibly resilient, with many meetups taking place virtually, several of which Go team members have been able to attend.

We’d like to help the entire open source community stay connected. In that vein, we’re excited to announce that Google will host a series of free virtual events, Google Open Source Live, every month through next year! As part of the series, on November 7th, members of the Go team will be sharing community updates, some things we’ve been up to, and a few best practices around getting started with Go.

Visit the official site for the Go Day on Google Open Source Live, to learn more about registration and speakers. To keep up-to-date with the Go team, make sure to follow the official Go twitter and visit, our hub for Go developers.

By Steve Francia – Product Lead, Go Team

Google Open Source Live: A monthly connection for open source communities

Tuesday, August 25, 2020

Starting in September, open source experts at Google will have a new place to meet with you online: Google Open Source Live, a virtual event series to connect with open source communities with a focus on different technologies and areas of expertise. Google Open Source Live launches on September 3, 2020, and will provide monthly content for open source developers at all levels, contributors, and community members. 
The inaugural event of this series will be: The new open source: Leadership, contributions and sustainability, in which the Google Open Source Programs Office, together with Developer Relations specialists, will share an overview of the best ways to get involved and succeed in the open source ecosystem with four exciting sessions.

Given how the 2020 pandemic has affected the communities’s ability to stay engaged and connect, it is important to us to stay present in the ecosystem. Therefore, we made a conscious decision to build an event series for developers to have the opportunity to hear directly from the Google Open Source Programs Office, developer advocates and experts. Each day will provide impactful information in a 2-hour time frame.

Attendee Experience

After attending several virtual events throughout the Summer, we designed our platform with one idea in mind: to create an alternative platform for developers to gather, learn, and interact with experts, and have fun.

Attendees can interact with the experts and speakers with Live Q&A chat during the sessions, and join an after party following the event! It’ll provide a great interactive opportunity for activities and to connect with others.

Sept. 3 Agenda

“The New Open Source: Leadership, contributions and sustainability”
9 AM - 11 AM PST




Hosted by Stephen Fluin, DevRel Lead and Dustin Ingram, Developer Advocate.  


"Be the leader you want in OSS"

Megan Byrd-Sanicki

Manager, Operations & Research


"5 simple things you can do to improve OSS docs"

Erin McKean, Docs Advocacy Program Manager, OSPO


Fireside Chat: "Business models and contributor engagement in OS"

Seth Vargo, Developer Advocate

Kaslin Fields, Developer Advocate


"Sustainability in OS"

Megan Byrd-Sanicki, Manager

 Operations & Research

Google Open Source Live Event Calendar

Each month will focus on one open source project or concept and feature several speakers who are subject matter experts in their fields. Events take place monthly on the first Thursday.



Sep 3

Oct 1

Nov 5

Dec 3

The new open source:

Leadership, contributions and sustainability

Knative day

On Google Open Source Live

Go day

On Google Open Source Live

Kubernetes day

On Google Open Source Live


Feb 4

Mar 4

Apr 1

May 6

Istio day

On Google Open Source Live

Bazel day

On Google Open Source Live

Beam day

On Google Open Source Live

Spark day

On Google Open Source Live

Jun 3

Jul 1

Aug 5

Sep 2

CDAP day

On Google Open Source Live

Airflow day

On Google Open Source Live

OSS Security day

On Google Open Source Live


Find out more 

Sign up to receive more details and alerts, and follow GoogleOSS@ and #GoogleOSlive for updates on Twitter.

By Jamie Rachel, Event Program Manager for the Google Open Source Programs Office

Assess the security of Cloud deployments with InSpec for GCP

Thursday, August 20, 2020

InSpec-GCP version 1.0 is now generally available, and two new Chef InSpec™ profiles have been released under an open source software license. The InSpec profiles contain controls for the GCP Center for Internet Security (CIS) Benchmark version 1.1.0 and the Payment Card Industry Data Security Standard (PCI DSS) version 3.2.1.

The Cloud Security Challenge

Developers are embracing automated continuous integration and continuous delivery (CI/CD), committing many application and infrastructure changes frequently. But centralized security teams can't review every application and infrastructure change. Those teams might have to block deployments (which decreases velocity and undermines continuous delivery) or review changes in production, where misconfigurations are more harmful and changes are more expensive.

Security reviews need to "shift left,” earlier in the software development lifecycle. Security teams likewise need to shift their own efforts to defining policies and providing tools to automate how compliance is verified. When developers adopt these tools, security and compliance checks become part of CI/CD, in a similar fashion to unit, functional, and integration tests, and thus become a normal part of the development workflow. Empowering developers to participate in this process means organizations can achieve continuous compliance. This also reinforces the mindset that security is everyone's responsibility.

What is InSpec

InSpec is a popular DevSecOps framework that checks the configuration state of resources in virtual machines and containers, on cloud providers such as GCP, AWS, and Azure. InSpec's lightweight nature, approachable domain-specific language, and extensibility make it a valuable tool for:
  • Expressing compliance policies as code
  • Enabling development teams to add tests that assess their applications' compliance with security policies before pushing changes to build and release pipelines
  • Automating compliance verification in CI/CD pipelines and as part of the release process
  • Unifying compliance assessments across multiple cloud providers and on-premises environments

InSpec for GCP and compliance profiles

The InSpec GCP resource pack 1.0 provides a consistent way to audit GCP resources. This release unifies the user experience by adding consistent behavior between resources and documentation for available fields. This resource pack also adds support for GCP endpoints that let you audit fields that are in beta (for example, GKE cluster pod security policy configuration).

You can use the GCP CIS Benchmark and the PCI DSS InSpec profiles to assess compliance with CIS and PCI DSS policies. CIS Benchmarks are configuration guides used by governments, businesses, industry, and academia. We strongly recommend configuring the workloads to meet or exceed these standards. PCI DSS is required for all organizations that accept or process credit card payments. The Terraform PCI Starter, coupled with the PCI InSpec profile, allows deployment of PCI-compliant environments and verifies their ongoing compliance.

This work is released under an open source license and we look forward to your feedback and contributions.

Validating PCI DSS and CIS compliance in infrastructure build pipelines

You can use InSpec to validate infrastructure deployments for compliance with standards such as PCI DSS and CIS. An automated validation process of new builds is important to detect insecure and non-compliant configurations as early as possible while minimizing the impact on developer agility.

With Cloud Build you can create CI pipelines for infrastructure-as-code deployments. You can run InSpec as an additional build step against resources in the GCP project to detect compliance violations in the target infrastructure. While this method doesn't prevent non-compliant build configurations, it does detect compliance issues, fail the build execution, and log the error in Cloud Logging. Cloud Build publishes build messages to a Cloud Pub/Sub topic, which can trigger a Cloud Function to integrate with appropriate alerting systems in case of a failed build. To prevent non-compliant infrastructure in a production environment, run the pipeline in a staging environment before promoting the content to production.

Here is an example pipeline definition for Cloud Build, using InSpec, to validate a project against the PCI guidelines. To run the PCI profile from a container inside a Cloud Build pipeline, clone the Git repository Payment Card Industry Data Security Standard (PCI DSS) version 3.2.1, build the Docker container from the root directory of the repository using the Dockerfile, and push the image to the Google Container Registry. The Cloud Build pipeline will store InSpec reports in a predefined bucket in json and html formats.

Here's an example for executing the PCI DSS InSpec profile as a step in a Cloud Build pipeline:

#...Previous execution steps
- id: 'Run PCI Profile on in-scope project'
  waitFor: ['Write InSpec input file']
  entrypoint: '/bin/sh'
    - '-c'
    - |
      inspec exec /share/. -t gcp:// \
      --input-file /workspace/inputs.yml \
      --reporter cli json:/workspace/pci_report.json \
      html:/workspace/pci_report.html | tee out.json

Note that in this example a previous execution step writes all required input parameters into the file /workspace/inputs.yml to make them available to the InSpec run. A CI/CD pipeline has been implemented for the PCI-GKE-Blueprint using Cloud Build and can be referenced as an example.

Try it yourself

Ready to try InSpec? Use this Cloud Shell Walkthrough to quickly install InSpec in your Cloud Shell instance and scan infrastructure in your GCP projects against the CIS Benchmark:

Chances are that in the walkthrough the InSpec scan detected some misconfigurations in your project.

As a developer of the project, you now know how to quickly scan your deployments, and you can begin to learn more about configuring your resources securely. Our Cloud Foundation Toolkit provides Terraform and Deployment Manager templates for best-practice configurations of your projects and underlying resources.

Most large organizations have platform teams that can adopt our Cloud Foundation Toolkit templates, which automate well-configured resource provisioning, and make those available to their developers. These organizations can also include InSpec testing steps in their CI/CD pipelines to provide early feedback to developers and to prevent misconfigured resources from getting released to Production.

By Bakh Inamov – Security and Compliance Specialist Engineer, Sam Levenick – Software Engineer, and Konrad Schieban – Infrastructure Cloud Consultant

Cloud Spanner Emulator Reaches 1.0 Milestone!

Wednesday, August 19, 2020

The Cloud Spanner emulator provides application developers with the full set of APIs, including the full breadth of SQL and DDL features that can be run locally for prototyping, development and testing. This offline emulator is free and improves developer productivity for customers. Today, we are happy to announce that Cloud Spanner emulator is generally available (GA) with support for Partitioned APIs, Cloud Spanner client libraries, and SQL features.

Since Cloud Spanner emulator’s beta launch in April, 2020, we have seen strong adoption of the local emulator from customers of Cloud Spanner. Several new and existing customers adopted the emulator in their development & continuous test pipelines. They noticed significant improvements in developer productivity, speed of test execution, and error-free applications deployed to production. We also added several features in this release based on the valuable feedback we received from beta users. The full list of features is documented in the GitHub readme.

Partition APIs

When reading or querying large amounts of data from Cloud Spanner, it can be useful to divide the query into smaller pieces, or partitions, and use multiple machines to fetch the partitions in parallel. The emulator now supports Partition Read, Partition Query, and Partition DML APIs.

Cloud Spanner client libraries

With the GA launch, the latest versions of all the Cloud Spanner client libraries support the emulator. We have added support for C#, Node.js, PHP, Python, Ruby client libraries and the Cloud Spanner JDBC driver. This is in addition to C++, Go and Java client libraries that were already supported with the beta launch. Be sure to check out the minimum version for each of the client libraries that support the emulator.

Use the Getting Started guides to try the emulator with the client library of your choice.

SQL features

Emulator now supports the full set of SQL features provided by Cloud Spanner. Some of the notable additions being support for SQL functions JSON_VALUE, JSON_QUERY, CEILING, POWER, CHARACTER_LENGTH, and FORMAT. We now also support untyped parameter bindings in SQL statements which are used by our client libraries written in languages with dynamic typing e.g., Python, PHP, Node.js and Ruby.

Using Emulator in CI/CD pipelines

You may now point the majority of your existing CI/CD to the Cloud Spanner emulator instead of a real Cloud Spanner instance brought up on GCP. This will save you both cost and time, since an emulator instance comes up instantly and is free to use!

What’s even better is that you can bring up multiple instances in a single execution of the emulator, and of course multiple databases. Thus, tests that interact with a Cloud Spanner database can now run in parallel since each of them can have their own database, making tests hermetic. This can reduce flakiness in unit tests and reduce the number of bugs that can make their way to continuous integration tests or to production.

In case your existing CI/CD architecture assumes the existence of a Cloud Spanner test instance and/or test database against which the tests run, you can achieve similar functionality with the emulator as well. Note that the emulator doesn’t come up with a default instance or a default database as we expect users to create instances and databases as required in their tests for hermeticity as explained above. Below are two examples of how you can bring up an emulator with a default instance or database: 1) By using a docker image or 2) Programmatically.

Starting Emulator from Docker

The emulator can be started using Docker on Linux, MacOS, and Windows. As a prerequisite, you would need to install Docker on your system. To bring up an emulator with a default database/instance, you can execute a shell script in your docker file to do so. Such a script would make RPC calls to CreateInstance and CreateDatabase after bringing up the emulator server. You can also look at this example on how to put this together when using docker.

Run Emulator Programmatically

You can bring up the emulator binary in the same process as your test program. Then you can then create a default instance/database in your ‘Setup’ and clean up the same when the tests are over. Note that the exact procedure for bringing up an ‘in-process’ service may vary with the client library language and platform of your choice.

Other alternatives to start the emulator, including pre-built linux binaries, are listed here.
Try it now

Learn more about Google Cloud Spanner emulator and try it out now.

By Asheesh Agrawal, Google Open Source

DEFCON Differential Privacy Training Launch

Tuesday, August 18, 2020

Differential privacy is a technique that enables organizations to learn from the majority of their data while simultaneously ensuring those results do not allow an individual’s data to be distinguished or re-identified. A popular way of attaining differential privacy is by adding noise to the data, which provides mathematical bounds on the amount of information that is leaked. Our open source offering aims to help developers implement differential privacy.

In the summer of 2019, we publicly launched our Differential Privacy Library. Since then, we’ve expanded it from just C++ to also include Go and Java.

We’ve come to realize that differential privacy requires more than just the library to be effectively implemented. We mentioned in a post earlier this summer that we want all developers to be able to interact with differential privacy, which requires more than an open-sourced library, but rather a training on the topic to share knowledge with all developers.

Our goal with this training is to provide a head start that is helpful for those considering differential privacy implementation. We also want to provide an experience on privacy and security that is understood and impactful to any individual in the field, whether they are a beginner or someone who has background knowledge in privacy.

This new training contains several steps and covers many topics, such as:
  • The foundations of differential privacy
  • Explanations as to why aggregation by itself may not hedge against privacy risks
  • The mathematical behind-the-scenes of noise
  • Tools that can be used in conjunction with differential privacy
  • Codelabs that users can take (in Go)
  • Additional resources to address any further questions

Step 1: Take our survey! It only takes five minutes!

This survey enables us to gain insights into what you are expecting to gain from this training. We are curious about what your objectives and goals are with this training, and if you have any experience with differential privacy.

Step 2: Check-out an introductory video to Differential Privacy!

We introduce topics like data aggregation, k-anonymity, differential privacy, noise, and others. The goal of this module is to introduce the foundations behind the differential privacy, and why it is an important and useful privacy tool.

Step 3: Try-out our codelabs

We have provided Codelabs in Go to help you practice implementing Differential Privacy library end-to-end.

Step 4: Learn more about differential privacy.

We want to offer an additional resource to help answer any questions you may have. If you have other resources that you find, please let us know and we will add these links to our overall training.

Step 5: Provide us with some feedback

Please use this survey as a platform to share your experience with this pilot. Did the content meet your expectations? Did it make sense? What was missing? This is the time for you to share your point of view and any pain points you experienced (as well as any positive aspects you encountered).

We hope this training provides an impactful experience from beginner coders to privacy specialists. The public differential privacy training will launch at the Stanford Biodesign: “Building for Digital Health” Buildathon, Sept 11-13, 2020, led by Stanford, and supported by Google Cloud and Apple Health engineers.

Please continue to reach out to us to share your experiences with us at The suggestions we receive will help us improve and it will inform our thinking as we add new features and updates.

Acknowledgements: Miguel Guevara, Bryant Gipson, Royce Wilson, Kate Frankenberg, Katie Holzheimer, Lior Gottleib, Carmen Bush

By Aditi Joshi – Security and Privacy Engineering, Google Cloud

Season of Docs announces 2020 technical writing projects

Monday, August 17, 2020

Season of Docs has announced the technical writers participating in the program and their projects! You can view a list of organizations and technical writing projects on the website.

The program received over 500 technical writer applications, and with them, over 800 technical writing project proposals. The enthusiasm from the technical writing and open source communities has been amazing!

What is next?

During the community bonding period from August 17 to September 13, mentors must work with the technical writers to prepare them for the doc development phase. By the end of community bonding, the technical writer should be familiar with the open source project and community, understand the product as a whole, establish communication channels with the mentoring organization, and set clear goals and expectations for the project. These are critical to the successful completion of the technical writing project.

Documentation development begins on September 14, 2020.

What is Season of Docs?

Documentation is essential to the adoption of open source projects as well as to the success of their communities. Season of Docs brings together technical writers and open source projects to foster collaboration and improve documentation in the open source space. You can find out more about the program on the introduction page of the website.

During the program, technical writers spend a few months working closely with an open source community. They bring their technical writing expertise to the project's documentation and, at the same time, learn about the open source project and new technologies.

The open source projects work with the technical writers to improve the project's documentation and processes. Together, they may choose to build a new documentation set, redesign the existing docs, or improve and document the project's contribution procedures and onboarding experience.

General timeline
August 16Google announces the accepted technical writer projects
August 17 - September 13Community bonding: Technical writers get to know mentors and the open source community, and refine their projects in collaboration with their mentors
September 14 - December 5Technical writers work with open source mentors on the accepted projects, and submit their work at the end of the period
January 6, 2021Google publishes the list of successfully-completed projects

See the full timeline for details, including the provision for projects that run longer than three months.

Find out more

Explore the Season of Docs website at to learn more about the program. Use our logo and other promotional resources to spread the word. Check out the FAQ for further questions!

By Kassandra Dhillon and Erin McKean, Program Managers, Google Open Source Programs Office

Java zPages for OpenTelemetry

Friday, August 14, 2020

What is OpenTelemetry?

OpenTelemetry is an open source project aimed at improving the observability of our applications. It is a collection of cloud monitoring libraries and services for capturing distributed traces and metrics and integrates naturally with external observability tools, such as Prometheus and Zipkin. As of now, OpenTelemetry is in its beta stage and supports a few different languages.

What are zPages?

zPages are a set of dynamically generated HTML web pages that display trace and metrics data from the running application. The term zPages was coined at Google, where similar pages are used to view basic diagnostic data from a particular host or service. For our project, we built the Java /tracez and /traceconfigz zPages, which focus on collecting and displaying trace spans.


The /tracez zPage displays span data from the instrumented application. Spans are split into two groups: spans that are still running and spans that have completed.


The /traceconfigz zPage displays the currently active tracing configuration and allows users to change the tracing parameters. Examples of such parameters include the sampling probability and the maximum number of attributes.

Using the zPages

This section describes how to start and use the Java zPages.

Add the dependencies to your project

First, you need to add OpenTelemetry as a dependency to your Java application.


For Maven, add the following to your pom.xml file:
        <artifactId>opentelemetry-sdk-extension-    zpages</artifactId>


For Gradle, add the following to your build.gradle dependencies:
implementation 'io.opentelemetry:opentelemetry-api:0.7.0'
implementation 'io.opentelemetry:opentelemetry-sdk:0.7.0'
implementation 'io.opentelemetry:opentelemetry-sdk-extension-zpages:0.7.0'

Register the zPages

To set-up the zPages, simply call startHttpServerAndRegisterAllPages(int port) from the ZPageServer class in your main function:
import io.opentelemetry.sdk.extensions.zpages.ZPageServer;

public class MyMainClass {
    public static void main(String[] args) throws Exception {
        // ... do work
Note that the package is required to use the default zPages setup. Please make sure your version of the JDK includes this package if you plan to use the default server.

Alternatively, you can call registerAllPagesToHttpServer(HttpServer server) to register the zPages to a shared server:
import io.opentelemetry.sdk.extensions.zpages.ZPageServer;

public class MyMainClass {
    public static void main(String[] args) throws Exception {
        HttpServer server = HttpServer.create(new                     InetSocketAddress(8000), 10);
        // ... do work

Access the zPages

View all available zPages on the index page

The index page (at /) lists all available zPages with a link and description.

View trace spans on the /tracez zPage

The /tracez zPage displays information about running and completed spans, with completed spans further organized into latency and error buckets. The data is aggregated into a summary-level table:

You can click on each of the counts in the table cells to access the corresponding span details. For example, here are the details of the ChildSpan latency sample (row 1, col 4):

View and update the tracing configuration on the /traceconfigz zPage.

The /traceconfigz zPage provides an interface for users to modify the current tracing parameters:


This section goes into the underlying design of our code.


The frontend consists of two main parts: HttpHandler and HttpServer. The HttpHandler is responsible for rendering the HTML content, with each zPage implementing its own ZPageHandler. The HttpServer, on the other hand, is responsible for listening to incoming requests, obtaining the requested data, and then invoking the aforementioned ZPageHandlers. The HttpServer class from is used to construct the default server and to handle http requests on different routes.


The backend consists of two components as well: SpanProcessor and DataAggregator. The SpanProcessor watches the lifecycle of each span, invoking functions each time a span starts or ends. The DataAggregator, on the other hand, restructures the data from the SpanProcessor into an accessible format for the frontend to display. The class constructor requires a TracezSpanProcessor instance, so that the TracezDataAggregator class can access the spans collected by a specific TracezSpanProcessor. The frontend only needs to call functions in the DataAggregator to obtain information required for the web page.


We hope that this blog post has given you a little insight into the development and use cases of OpenTelemetry’s Java zPages. The zPages themselves are lightweight performance monitoring tools that allow users to troubleshoot and better understand their applications. Once OpenTelemetry is officially released, we hope that you try out and use the /tracez and /traceconfigz zPages!

By William Hu and Terry Wang – Software Engineering Interns, Core Compute Observability

Google Summer of Code 2020 Statistics: Part 2

Thursday, August 13, 2020

With the program nearing the end of the summer, it’s time for another round of updates!


The 1,198 students accepted into the GSoC 2020 program came from 550 universities, of which, 114 have students participating for the first time in GSoC.

Schools with the most accepted students for GSoC 2020:
University# of Accepted Students
Indian Institute of Technology, Roorkee48
Indian Institute of Technology, Kanpur27
International Institute of Information Technology, Hyderabad24
National Institute of Technology Karnataka, Surathkal23
Birla Institute of Technology and Science, Pilani (BITS Pilani)13
Indian Institute of Technology, Kharagpur13
Indian Institute of Technology (BHU), Varanasi11
University of Moratuwa11
National Institute of Technology, Hamirpur10
Amrita Vishwa Vidyapeetham, Amritapuri Campus10
University of Tokyo10
University Of Colombo School Of Computing (UCSC)10


Each year we pore over gobs of data to extract some interesting statistics about the GSoC mentors. Here’s a quick synopsis of our 2020 crew:
  • Registered mentors: 3,592
  • Mentors with assigned student projects: 2,156
  • Mentors who have participated in GSoC for 10 or more years: 78
  • Mentors who have been a part of GSoC for 5 years or more: 199
  • Mentors that are former GSoC students: 533 (24.7%)
  • Mentors that have also been involved in the Google Code-in program: 405 (18.8%)
  • Percentage of new mentors: 34.18%
GSoC 2020 had an international representation with mentors from 67 countries around the world!

The global pandemic, COVID-19, brought additional challenges to this year’s GSoC program. Whether living with the virus, adjusting to shifting school and work schedules, or pivoting to a remote lifestyle, students and mentors have had to prioritize their safety and delicately balance their new way of life. Despite these unprecedented times, our students continue to push on and our mentors fully support our students by sharing their passion for open source, listening to their concerns and providing them with valuable advice. For that commitment, we would like to acknowledge and give thanks to all students and mentors in the GSoC 2020 program. Not even a pandemic can dampen your enthusiasm and tireless contributions to the open source community!

By Stephanie Taylor – Program Manager, Google Open Source Programs Office

Supporting Wikipedia with more tools for editors

Tuesday, August 11, 2020

Google has a commitment to making information more accessible to people around the world, while ensuring that the information on the web is accurate and reflects the diversity of its users. Just like Google supports open source software development, WikiLoop is one of the explorations on better contributing to the open knowledge movement. To this day, Wikipedia, a Wikimedia movement project, remains a reliable information source that is also available with an open license, which makes it possible for Knowledge Engine of Google—as well as knowledge graph systems operated by others —to draw excerpts from it for Search features and other apps. With the project’s sustainability in mind, Google has contributed back to the Wikimedia movement in a number of ways since 2018. Building on this commitment, Google created the umbrella program WikiLoop in 2019, which hosts several tools for editors that focus on content quality, like WikiLoop DoubleCheck. 
WikiLoop is led by Zainan Zhou—a Googler for the last 7 years, and a Wikipedian for the last 5—who works as a software engineer in the Knowledge Engine team at Google. When he joined the free encyclopedia as an editor, he always wondered how his company could contribute to this open source project. Zainan involved the Wikipedia communities in every step of the development of WikiLoop, connecting with editors from different parts of the world at Wikimedia events, like WikidataCon, Wikimania, Wiki Conference North America, and WikiDevSummit, throughout 2019. The most recent involvement with the community of Wikipedia editors included a consultation and vote to change the name of the most popular WikiLoop artifact, a tool for peer-review of Wikipedia articles, DoubleCheck.

In the past few months, we focused on raising awareness of WikiLoop DoubleCheck. This tool allows registered and unregistered users to mark new edits with tags “looks good”, “not sure”, and “should revert”, a peer review system which editors could use to approve or revert new content on Wikipedia. Since its launch, the tool has witnessed a 309% quarter over quarter growth in tags added, and over 1,000 editors have used it to review Wikipedia content. With the help of volunteer translators and machine translation, WikiLoop DoubleCheck is now made available in 25 languages, and we hope to continue serving more Wikipedia editors in the months to come. In order for Google’s Knowledge Engine to organize the world's information, the knowledge source needs to be healthy. While peer-review on Wikipedia is an established process that has been going on for years, tools like WikiLoop DoubleCheck support the thousands of volunteers who dedicate their time to this task on Wikipedia by making information verification more accessible.

The WikiLoop program was originally conceived as a virtuous circle: providing data and tools to enhance human editor's productivity, and making the Wikipedia editorial input more machine-readable for open knowledge institutions, academia and researchers interested in advancing machine learning technology.

WikiLoop leverages Google’s talents at software development to contribute to global Wikipedia content accuracy by enhancing the existing suite of Wikimedia and community tools for content validation at scale. While WikiLoop is a contribution of Google to the Wikipedia communities, Google and the Wikimedia Foundation have partnered in other areas as well. Learn more about Google’s partnership with the Wikimedia Foundation on the partnership’s page on Meta-Wikimedia

While probably the most popular in the set, DoubleCheck is not the only tool under the WikiLoop umbrella. We are also building data sets and tools and continue to explore other opportunities to contribute to the open knowledge movement. Learn more about tools and other initiatives, like the Coalition call, on the program’s page on Meta-Wikimedia.

By María Cruz  Program Manager, Google Open Source Programs Office

Introducing TensorFlow Recorder

Friday, August 7, 2020

When training computer vision machine learning models, data loading can often be a performance bottleneck, causing your GPU or TPU resources to be underutilized while waiting for data to be loaded into the model. Storing your dataset in the efficient TensorFlow Record (TFRecord) format is a great way to solve these problems, but creating TFRecords can unfortunately often require a great deal of complex code.

Last week we open sourced the TensorFlow Recorder project (also known as TFRecorder), which makes it possible for data scientists, data engineers, or AI/ML engineers to create image based TFRecords with just a few lines of code. Using TFRecords is incredibly important for creating efficient TensorFlow ML pipelines, but until now they haven’t been so easy to create. Before TFRecorder, in order to create TFRecords at scale you would have had to write a data pipeline that parsed your structured data, loaded images from storage, and serialized the results into the TFRecord format. TFRecorder allows you to write TFRecords directly from a Pandas dataframe or CSV without writing any complicated code.

You can see an example of TFRecoder below, but first let’s talk about some of the specific advantages of TFRecords.

How TFRecords Can Help

Using the TFRecord file format allows you to store your data in sets of files, each containing a sequence of protocol buffers serialized as a binary record that can be read very efficiently, which will help reduce the data loading bottleneck mentioned above.

Data loading performance can be further improved by implementing prefetching and parallel interleave along with using the TFRecord format. Prefetching reduces the time of each model training step(s) by fetching the data for the next training step while your model is executing training on the current step. Parallel interleave allows you to read from multiple TFRecords shards (pieces of a TFRecord file) and apply preprocessing of those interleaved data streams. This reduces the latency required to read a training batch and is especially helpful when reading data from the network.

Using TensorFlow Recorder

Creating a TFRecord using TFRecorder requires only a few lines of code. Here’s how it works.
import pandas as pd
import tfrecorder
df = pd.read_csv(...)

TFRecorder currently expects data to be in the same format as Google AutoML Vision.

This format looks like a pandas dataframe or CSV formatted as:

  • split can take on the values TRAIN, VALIDATION, and TEST
  • image_uri specifies a local or google cloud storage location for the image file.
  • label can be either a text-based label that will be integerized or an integer
In the future, we hope to extend TensorFlow Recorder to work with data in any format.

While this example would work well to convert a few thousand images into TFRecords, it probably wouldn’t scale well if you have millions of images. To scale up to huge datasets, TensorFlow Recorder provides connectivity with Google Cloud Dataflow, which is a serverless Apache Beam pipeline runner. Scaling up to DataFlow requires only a little bit more configuration.

What’s next?

We’d love for you to try out TensorFlow Recorder. You can get it from GitHub or simply pip install tfrecorder. Tensorflow Recorder is very new and we’d greatly appreciate your feedback, suggestions, and pull requests.

By Mike Bernico and Carlos Ezequiel, Google Cloud AI Engineers

Open source by the numbers at Google

Wednesday, August 5, 2020

At Google, open source is at the core of our infrastructure, processes, and culture. As such, participation in these communities is vital to our productivity. Within OSPO (Open Source Programs Office), our mission is to bring the value of open source to Google and the resources of Google to open source. To ensure our actions match our commitment, in this post we will explore a variety of metrics intended to increase context, transparency, and accountability across all of the communities we engage with.

Why we contribute: Open source has become a pervasive component in modern software development, and Google is no exception. We use thousands of open source projects across our internal infrastructure and products. As participants in the ecosystem, our intentions are twofold: give back to the communities we depend on as well as expand support for open source overall. We firmly believe in open source and its ability to bring together users, contributors, and companies alike to deliver better software.

The majority of Google’s open source work is done within one of two hosting platforms: GitHub and git-on-borg, Google’s production Git service which integrates with Gerrit for code review and access control. While we also allow individual usage of Bitbucket, GitLab, Launchpad, and other platforms, this analysis will focus on GitHub and git-on-borg. We will continue to explore how best to incorporate activity across additional channels.

A little context about the numbers you’ll read below:
  • Business and personal: While git-on-borg hosts both internal and external Google created repos, GitHub is a mixture of Google projects, experimental efforts and personal projects created by Googlers.
  • Driven by humans: We have created many automated bots and systems that can propose changes on both hosting platforms. We have intentionally filtered these data to ensure we are only showing human initiated activities.
  • GitHub data: We are using GH Archive as the primary source for GitHub data, which is currently available as a public dataset on BigQuery. Google activity within GitHub is identified by self registered accounts, which we anticipate under reports actual usage as employees acclimate to our policies.
  • Active counts: Where possible, we will show ‘active users’ and ‘active repositories’ defined by logged activity within each specified timeframe (for GH archive data, that’s any event type logged in the public GitHub event stream).
As numbers mean nothing without scale, let’s start by defining our applicable community: In 2019, more than 9% of Alphabet’s full time employees actively contributed to public repositories on git-on-borg and GitHub. While single digit, this percentage represents a portion of all full time Alphabet employees—from engineers to marketers to admins, across every business unit in Alphabet—and does not include those who contribute to open source projects outside of code. As our population has grown, so has our registered contributor base:
This chart shows the aggregate per year counts of Googlers active on public repositories hosted on GitHub and git-on-borg

What we create: As mentioned above, our contributing population works across a variety of Google, personal, and external repositories. Over the years, Google has released thousands of open source projects (many of which span multiple repositories) and ~2,600 are still active. Today, Google hosts over 8,000 public repositories on GitHub and more than 1,000 public repositories on git-on-borg. Over the last five years, we have doubled the number of public repos, growing our footprint by an average of 25% per year.

What we work on: In addition to our own repositories, we contribute to a wide pool of external projects. In 2019, Googlers were active in over 70,000 repositories on GitHub, pushing commits and/or opening pull requests on over 40,000 repositories. Note that more than 75% of the repos with Googler-opened pull requests were outside of Google-managed organizations (on GitHub).
This charts shows per year counts of activities initiated by Googlers on GitHub

What we contribute: For contribution volume on GitHub, we chose to focus on push events, opened, and merged pull requests instead of commits as this metric on its own is difficult to contextualize. Note that push events and pull requests typically include one or more commits per event. In 2019, Googlers created over 570,000 issues comments, opened over 150,000 pull requests, and created more than 36,000 push events on GitHub. Since 2015, we have doubled our annual counts of issues created and push events, and more than tripled the number of opened pull requests. Over the last five years, more than 80% of pull requests opened by Googlers have been closed and merged into active repositories.

How we spend our time: Combining these two classes of metrics—contributions and repos—provides context on how our contributors focus their time. On GitHub: in 2015, about 40% of our opened pull requests were concentrated in just 25 repositories. However, over the next four years, our activity became more distributed across a larger set of projects, with the top 25 repos claiming about 20% of opened pull requests in 2019. For us, this indicates a healthy expansion and diversification of interests, especially given that this activity represents both Google, as well as a community of contributors that happen to work at Google.
This chart splits the total per year counts of Googler created pull requests on GitHub by Top 25 repos vs the remainder ranked by number of opened pull requests per repo per year.

Open source contribution is about more than code

Every day, Google relies on the health and continuing availability of open source, and as such we actively invest in the security and sustainability of open source and its supply chain in three key areas:
  • Security: In addition to building security projects like OpenTitan and gVisor, Google’s OSS-Fuzz project aims to help other projects identify programming errors in software. As of the end of 2019, OSS-Fuzz had over 250 projects using the project, filed over 16,000 bugs, including 3,500 security vulnerabilities.
  • Community: Open source projects depend on communities of diverse individuals. We are committed to improving community sustainability and growth with programs like Google Summer of Code and Season of Docs. Over the last 15 years, about 15,000 students from over 105 countries have participated in Google Summer of Code, along with 25,000 mentors in more than 115 countries working on more than 680 open source projects.
  • Research: At the end of 2019, Google invested $1 million in open source research, partnering with researchers at UVM, with the goal to deepen understanding of how people, teams and organizations thrive in technology-rich settings, especially in open-source projects and communities.
Learn more about our open source initiatives at

By Sophia Vargas – Researcher, Google Open Source Programs Office