Category: Monitoring

I recently had an issue with a host that was already months ago added to Dynatrace, but needed to be readded. I didn’t show up in the OneAgent deployment status screen.

Normally, I would be successful by restarting the OneAgent service on Dynatrace on the host I wanted to add.

This is a quick snippet, a view onto a Statuscake response time graph. A web application had a memory leak, and over multiple days, the response time was building up on instance 1. If instance 2 is queried, the response time is good; instance 1 had the memory leak and would cause the response time to build up.

There is nothing to learn here, but this graph speaks for itself and looks rather beautiful. It shows an extremely rare and very visible error.

If you’re considering monitoring your website using Statuscake, there are two potential issues you should be aware of. Firstly, Statuscake relies on virtual machine hoster on the internet, and their IP ranges may be impacted by DDoS mitigation efforts. This means that individuals from the same ASN or subnets may be involved in attacking your website. Once you start blocking providers of compute resources on the internet, your monitoring might start to become affected too and become flaky.

I’ve been recently exploring the monitoring and capturing of DNS entry history for various types of DNS records, with a focus on (SOA, NS, MX, AAAA, A, TXT, CNAME) records. When working with different companies, you might build services that rely on DNS operated by other companies and individuals. Mistakes can occur when the service goes offline or automatic TLS certificate issuance fails. It’s important to monitor DNS records, set up alerts, and maintain a historical record of changes for specific DNS entries, including the chain of CNAME resolutions, without excessively capturing the rotation of DNS entries.