jhuangtw / xg2xg Goto Github PK

Mainly used ACLaim, Sphinx and Buganizer for Identity and Access Mgmt. What would be analogous in the real world?

Please add one if anyone knows there is a similar service. The key feature is that anyone can create an mdb group, not by the admins.

https://chromium.googlesource.com/infra/goma/client/

Currently, it appears this repository is maintained by a single person. Come up with a system where multiple people are allowed to approve changes (e.g. prow from k8s).

https://bazel.build/ (sorry I'm on mobile, couldn't make a PR)

srcfsd (the other FUSE file system we use in our builds, which exposes the monolithic Google repository)

-- https://blog.bazel.build/2018/04/13/preliminary-sandboxfs-support.html

http://memegen.com/ is not a meme-generator of any sorts. Consider removing the row to avoid disappointing other Xooglers?

I joined Google in the last year and have been working with big data tools for a while. Spark should be considered comparable to Flume. Also, Apache Flink should be on the same level, comparable to Flume. Both Spark and Flink are relatively high-level interfaces that will build DAGs from what are effectively macros. They provide db-like operations like joins, filters, etc.

valentine is intended for groups, not individual use, so it's probably closer to Thycotic Secret Server or Hashicorp Vault rather than the alternatives currently listed (since they lack audit logs, ACLs, etc)

I've used GWP few times, mainly with Java, and I was amazed at the tech, today there was an annoucement from a company that does something similar to it

https://documentation.prodfiler.com/#what-is-prodfiler

To quote: Prodfiler is inspired by software available inside some Hyperscalers, most notably Google and Facebook. Google has an internal system called ‘Google-Wide Profiling’, which has led to very significant savings, a culture of performance awareness/improvement, and to long-term developments such as SwissTables and even the various hardware offload designs. Likewise, Facebook runs a system, which is described in this talk from the SPARK AI Summit.

Snippyly is an external version of Google's most used internal productivity tool, built by an ex-Googler.
It is a realtime visual collaboration tool: you can take screenshots of your screen, annotate and collaborate realtime (like google docs).

Website link:
https://snippyly.com/

Chrome store link:
https://chrome.google.com/webstore/detail/snippyly/okdocmoamfmoeiifjhmgklgndmhaeaio

https://cloud.google.com/bigquery/

Wondering if anyone has come across any linter that do a similar functionality?

Monarch: https://research.google/pubs/pub50652/

Napa: https://research.google/pubs/pub50617/

Both of them are public and used internally.

I am not sure what are external equivalents

Is there a good open-source equivalent to analog?

Is there an open source equivalent to piper, Google's version control tool? - Stack Overflow mentioned that Subversion should be sufficient for small companies, while large companies can use Perforce.

another protobuf alternative?
https://capnproto.org/

I see a lot of good entries but no description at a glance before checking the actuals link of real world :/

What are some open source tools that can be used to visualize data inside of a sstable equivalent or inside of a proto?

There was an excellent doc about capability vs ACL-based permissions, called "Why Dilbert can't ACL", but I can't find any equivalent blog post outside. I also don't know who was the author.

Does anybody have a pointer to the contents of this? :)

Could you please add an explicit LICENSE file to the repo so that it's clear under what terms the content is provided, and under what terms user contributions are licensed?

Per GitHub docs on licensing:

[...] without a license, the default copyright laws apply, meaning that you retain all rights to your source code and no one may reproduce, distribute, or create derivative works from your work. If you're creating an open source project, we strongly encourage you to include an open source license.

Since this is a text-only repo rather than code, maybe something like CC-BY-4.0 would be a good option? See the high-level overvivew for the summary.

Thanks!

Hi, this repo topic is too big, but the story inside is so limited, the things are almost known staff and open source projects, in my humble opinion, you just mentioned the alternatives, I would expect if u could share more detailed thoughts...

Is there an open-source equivalent to CDD?

Since CDD is pretty new within Google, it works like this:

• define a proto that is your config
• implement a stubby service provided by CDD
• on startup, register your proto with CDD

Then you can push config updates to CDD from the CLI and it will broadcast those changes to all your instances by calling the stubby endpoint. It's useful for updating config values without having to restart servers.

I find myself missing having a tool that can trace through the whole RPC call-graph for a given session. Really nice for debugging and understanding the broader system as a newbie by inspecting requests/responses.

It sounds like Datadog APM may be able to do similar things? Any limitations/differences to be aware of?

At a glance it seems like the use cases are a bit different. Datadog APM shows the latencies, but I can't find request/response payloads. Looks like it runs on every request, so it can't log everything. I didn't use Dapper much but maybe that's closer to what Datadog APM does - low-overhead performance monitoring.

xsds/sherlog was opt-in just for a given session and seemed to be geared at debugging specifically. It was great when you had just found a repro and could record a full trace to reference for anything unusual downstream. With Datadog APM it looks like I'd have to use the right filters to get to the right trace.

I've often wondered what the closest thing to Boq and Wiz would be in the outside world. Any ideas?

It`s interesting

Google's Manifold is a cross-platform, RPC-based job processing framework that I used a couple times within Google. Now I'm in the position of wanting something similar externally, but I'm having trouble spotting anything equivalent. Anything similar?

This was a very simple and powerful framework for building continuous pipelines at Google. I am not able to find anything remotely close to it. Any suggestions?