Operations | Monitoring | ITSM | DevOps | Cloud

C-suiter, VP’er, or anyone who’s top of the pile in an enterprise faces inherent pressure that comes part and parcel with the role that they’re taking on. Many of the pressures can’t be overcome, it’s simply the nature of the beast. But dealing with technical issues in their day-to-day life is one of their biggest gripes, because it always feels like a problem that should be solved – not one needing to be dealt with again and again.

In Java applications, concurrency issues can be difficult to reproduce and debug. Because work is scheduled nondeterministically across threads, the conditions that have led to an error in one execution of the program may not trigger the same issue the next time around. Exceptions that are silently handled—also known as swallowed exceptions—can also be challenging to debug because they typically do not leave any trace in the logs.

Like everyone else in the world, we are thinking hard about how we can harness the power of AI and machine learning while also staying true to our core values around respecting the security and privacy of our users’ data. If you use Sentry, you might have seen our “Suggested Fix” button which uses GPT-3.5 to try to explain and resolve a problem. We have additional ideas being developed as well that we’re excited to preview.

Orange España, Spain’s second largest mobile operator, suffered a major outage on January 3, 2024. The outage was unprecedented due to the use of RPKI, a mechanism designed to protect internet routing security, as a tool for denial of service. In this post, we dig into the outage and the unique manipulation of RPKI.

Imagine a symphony where every musician plays their part flawlessly, but without a conductor to guide the orchestra, the result is just a discordant mess. Now apply that image to the modern IT landscape, where development and operations teams work with remarkable autonomy, each expertly playing their part. Agile methodologies and DevOps practices have empowered teams to build and manage their services independently, resulting in an environment that accelerates innovation and development.

Deploying an infrastructure in the cloud has become THE standard across the industry. According to O’Reilley Research, more than 90% of organizations are on the cloud or are using it in some capacity. As reported by Precedence Research, cloud deployment will surpass $1 trillion by the year 2028. But with the adoption of a new platform comes the need for tools and best practices to keep it up and running efficiently, which includes cloud-based monitoring.

In the rapidly evolving world of Large Language Models (LLMs), ensuring peak performance and reliability is more critical than ever. This is where the concept of 'LLM Observability' comes into play. It's not just about monitoring outputs; it's about gaining deep insights into the internal workings of these complex systems.

If you want to bring down your monitoring costs, you need to shake up a decision paralysis in engineering.

In part 1, we looked at how the design plan of traditional monitoring technologies depended heavily on properties of the systems that were intended to monitor and then showed how those properties began to be undermined by an increase in complexity, an increase which can ultimately be captured by the concept of entropy. In this part, we will explore how increased entropy forces us to rethink what is required for monitoring.