Operations | Monitoring | ITSM | DevOps | Cloud

We are happy to announce the release of two new versions of Icinga 2 today, 2.16.0 and 2.15.3. The first one includes some new features highlighted below, as well as a number of bug fixes and other improvements. The latter one is a small bug fix release that brings some of the other fixes included in 2.16.0 to the 2.15.x branch as well.

Your LLM observability stack probably handles individual model calls well enough. Latency, token counts, error rates, maybe even evaluation scores....

The April 2026 update to UK Cyber Essentials marks an important shift. Not because it introduces radically new security concepts, but because it removes tolerance for inconsistency. With the effective date quickly approaching, many UK organizations are focused on meeting the immediate requirements. That matters. But the more durable story is what these changes reveal about how security and compliance are now expected to operate in real world environments.

In the wake of 2010s, PaaS felt like magic. You focused on the code, and the platform did the rest. You could ship a production app without knowing anything about networking or, heck, even what a load balancer is. Heroku in particular made deployment a lost thought, especially for early-stage companies. That era is somewhat over, not because platforms got worse overnight, but because the assumptions underneath them quietly stopped being true.

Transform your IT service management data into actionable insights with dashboards that give your team real-time visibility without the setup headache.

If your Apache ActiveMQ deployment has a growing ActiveMQ.DLQ, you are not alone, and you are looking at the right problem. An unbounded, unmonitored dead letter queue is one of the most common root causes of "invisible" message loss in enterprise messaging environments. DLQ messages land without fanfare, nobody notices, and business-critical data quietly disappears from the processing pipeline.

When engineers search for "Apache ActiveMQ vs Apache Artemis," most of what they find is either a shallow feature checklist or a confident recommendation to "just migrate to Apache Artemis." Neither helps a senior architect deciding whether to stay on a stable, battle-hardened Apache ActiveMQ deployment, or a platform team evaluating both options for a new system with clear eyes.

Part 2: Key insights from a fireside chat with Chris Yates. Read part 1 here. Choosing a database monitoring vendor isn't just about features. Once you’re confident that it’s time to reassess your database monitoring strategy, the natural instinct is to start comparing products. However, it’s vital to know how to assess vendor relationships, support quality, and product innovation before you sign anything.

SREs and platform teams are overwhelmed by the effort of manually troubleshooting ever-more complex cloud-native environments. This pain is driving a breakneck adoption of AI SRE solutions that promise to automate core reliability practices, from root cause analysis to capacity planning. For teams with strong engineering talent, creating a DIY AI SRE seems like a straightforward challenge.

AI has quickly become part of the language of network observability. Many vendors across the observability landscape can describe, summarize, correlate, or explain some data or situation, leveraging basic LLM capabilities. At a distance, many of these offerings sound similar. They promise faster insight, efficient operations, and a more intelligent path through rising complexity. But the industry has reached a point where surface-level similarity is creating noise, not value.