This is the second post in a series on common Go footguns. This post demonstrates the performance impacts of appending a known number of elements to a slice without doing any pre-allocation. This is one of the most common and easily fixed performance issues I see in pull requests on a regular basis.

Describing the common footgun of checking an interface for nil

Some projects make rather heavy use of gomock for testing, and I’ve noticed the command that I had been using to run all tests and generate coverage included the mocks in coverage and brought the numbers down.

Automatic backups of PostgreSQL hosted in kubernetes

A more advanced and useful multiwan failover setup. Block some VLANs from accessing the failover connection, limiting failover use to actual users or other high priority traffic.

A simple multiwan failover setup incase of primary ISP outage.

CRUD OPNsense Unbound DNS with CLI or External DNS

A previous post described how to monitor a Synology NAS with SNMP and Prometheus. This post goes over the Ansible role created to deploy the snmp_exporter. I’m rather dissatisfied with the way I’m doing testing with molecule at the moment, so I’m not releasing the full role in a repo at this time. However the important files are layed out. It would probably be nicer to build an image with the SNMP exporter and deploy a container, but this Ansible deployment was quick and easy to bang out.

My Synology NAS is a very important part of my home network, and until recently it was practically unmonitored. Synology doesn’t make it easy to monitor their NAS devices out of the box. It’s possible to get Prometheus’ node exporter running, but that doesn’t export all of the metrics of interest and is generally a pain to deal with. I also prefer to minimize the number of things running on my NAS.

Monitoring with Prometheus and Grafana at home