article --summary
The summary
The piece walks through steps for getting Horizon configured properly, starting with making timeout and retry_after values agree with each other. It also covers confirming that deploy hooks run as expected and gathering real measurements of how queues behave before deciding to add more workers. The author's central argument is that slow queues usually trace back to unclear configuration and missing visibility rather than a simple lack of capacity.
Editorial note: This is a summary of an article first published by freek.dev. The full piece is on their site.