Hi there 👋

Today we will explore how to leverage atomic operations without opening a multi-statement transaction, increasing the system’s write throughput without sacrificing correctness. To avoid conflicts with other processes changing the same data, we can lock the rows before updating them, for example with SELECT ... FOR UPDATE. This approach has drawbacks: locks consume database resources for the entire duration of the transaction and can lead to lock contention as writers wait for one another, reducing write throughput. It is generally preferable for transactions to be short. ...

Conflict-free Replicated Data Types and event-driven systems are closely related, let’s explore it.

Faster Elasticsearch tests without starting containers for each test.

Update documents in Elasticsearch without losing data.

Making better technical decisions with decision records.

A practical guide to issuing TLS certificates in Go: self-signed, Certificate Authority, Certificate Signing Request.

Learn how to use HTTP/3 DATAGRAMs in Go for unreliable data transmission.

Previous parts: Writing HTTP3 Server HTTP3 Client Server Stream Today we will create our own HTTP/3 client stream to send some sweet-sweet data to the server. Streaming data from the client can be used for sending large or real-time data uploads, such as file uploads, telemetry, or live feeds. With HTTP/3 and Go, you can efficiently stream request bodies without buffering the entire payload in memory. Streaming Client Request This is similar to streaming a request body in regular HTTP. We’ll use an io.Pipe to write data from the client as it becomes available, while the server reads the stream as it arrives. ...

Use golden files for testing APIs

Create desktop applications using Go and Webview, packaging them into a single executable.