Skip to content
shamimstack

P·04 — Maintainer

Overview

Queue depth, wait time and failure rate as first-class Prometheus metrics.

Documentation for QueueLens 1 min read

The problem

Horizon tells you how many jobs are waiting. It does not tell Prometheus, so a platform team either scrapes Redis itself or finds out about a queue backlog when a customer does.

What it is

QueueLens is a small exporter that reads Horizon's Redis structures and publishes queue depth, oldest-job age, throughput and failure counts as Prometheus metrics. A bundled Grafana dashboard turns those into a queue-depth SLO.

Design rule

Every metric names its queue and its connection. An aggregate that hides which queue is backed up is a dashboard that looks healthy while one worker starves.

Outcome

Four internal teams adopted it in the first quarter, and it now ships with the default platform image, so a new service gets queue visibility without adding an exporter.

Enter your credentials to continue to your workspace.
Email address
Remember me
Don't have an account? Sign up
Enter your details below to create your account
Name
Email address
Password
Confirm password
Already have an account? Log in
Enter your email and we will send you a reset link
Email address
Or, return to log in