Operations | Monitoring | ITSM | DevOps | Cloud

Exploratory data analysis is a key component of the machine learning pipeline that helps in understanding various aspects of a dataset. For example, you can learn about statistical properties, types of data, the presence of null values, the correlation among different variables, etc. But to get these details, you need to use different types of Python methods and write multiple lines of code.

As we continue adding to our growing catalog of more than 100 plugins for Grafana, we have been focused on developing data sources for Grafana that are more purpose-built for the respective technologies. One example has been the recent update to our core Prometheus data source. We have deprecated AWS authentication from the original Prometheus data source, and we created a new dedicated Amazon Managed Service for Prometheus plugin that will specifically cater to the AWS use case.

If you remember the last update, I originally planned to ship quite a few features in June. Life and family had to take priority for me, so they didn't get done. Thankfully, I'm building OnlineOrNot for the extremely long term, so there's no rush to ship as fast as possible. A fair few things did get shipped though.

At BugSplat, we've been supporting applications and video games with crash and error reporting for a long time. Over the years, we've collaborated with a wide range of teams, handling applications of all sizes. From our experience and numerous conversations with users, we've noticed an interesting trend: the distribution of crashes isn't uniform. If your application experiences 100 crashes in a given version, those crashes aren't caused by 100 different defects.

DevOps Research and Assessment (DORA) metrics are a ubiquitous measure of DevOps performance. These metrics are used in nearly every enterprise engaged in software development. DORA metrics help measure DevOps maturity, identify bottlenecks, and guide quality and process improvements. Despite their popularity, DORA metrics are generally considered difficult to measure and are primarily used by technical teams within the context of their respective domains.

Today we’re launching Fair Source, a new approach to software sharing that is safe for companies to adopt and developers to use. Before Fair Source, companies that wanted to engage the developer community with their core products often did not know how to do so while maintaining control over their roadmap and business model. The result is that most software products today are closed-source. With Fair Source, companies have a new option. The Fair Source option is not theoretical.

While retrospectives provide a valuable pathway for learning outside of the flow of work, we also want learning to happen during an incident or unexpected event as it unfolds. This can be challenging due to the negative impact of stress on our ability to learn and navigate difficult situations. In this article, we’ll dig into how stress inhibits our ability to learn and what we can do about it.