pkgsite VS memberlist

Reading the standard library will give you ideas/insight about various Go idiomatic patterns/approaches, and you can see a full website/API implementation in the pkg.go.dev repository (https://github.com/golang/pkgsite). Projects like https://github.com/etcd-io/etcd may be interesting too.

Let's look at some real codebase for an example on how to write database tests without mocking. The source code for the Go package discovery site(https://pkg.go.dev/) is available at[1] That site uses postgres as its primary database[2]. The database package has a method called GetLatestInfo[3] that fetches the latest versions of a module. That method is called from the frontend http handlers[4] via an interface[5] When it comes to testing that frontend handler, you would expect the tests to use a mock implementation of that interface method. But that's not what they do, instead they use a real postgres database in the test[6].

The code behind pkg.go.dev is also open-source and might be an interesting read.

BTW, the number of dependencies of pkgsite surprises me: https://github.com/golang/pkgsite/blob/master/go.mod

The source for Go's pkg site has been helpful to me https://github.com/golang/pkgsite

In my unpopular opinion, the go net/http is good enough to develop the web http application. The standard API is clean and well-documented. One of example is the pkgsite (https://github.com/golang/pkgsite).

It's also available as a mirror at https://github.com/golang/pkgsite. All the golang.org/x/* packages are thankfully available there, making them pretty easy to find.

I use lightweight https://github.com/hashicorp/memberlist to build initial cluster skeleton (find out what are nodes constituting it)

Does https://github.com/hashicorp/memberlist count? It’s a gossip-based eventual consistency protocol based on SWIM.

An equivalent of golang's memberlist would be awesome.

I don't know the engineer that did it but I know what happened. There's a library that a bunch of foundational Amazon services use called DFDD. It uses a gossip protocol to handle service discovery and health checking. The open source equivalent is Hashicorp memberlist. To remove a node from the cluster, you have to send a command to an arbitrary node that says a node is dead.

serf uses memberlist which uses the SWIM failure detection protocol (https://www.cs.cornell.edu/projects/Quicksilver/public_pdfs/SWIM.pdf) with the Lifeguard extensions (https://arxiv.org/pdf/1707.00788.pdf).

https://github.com/hashicorp/memberlist

Fairly idiomatic/clean

This algorithm assumes you're doing that separately using some sort of membership protocol (e.g. SWIM), and you pass in the membership list that you are keeping up to date with that protocol. If you're curious about those, HashiCorp memberlist and HashiCorp Serf are really good to look at. I believe Consul uses those libraries under the hood.