Grafana Dazzle at ObservabilityCon

and New Relic claim an AI first

Welcome!

Welcome to the fifth edition of the Observability 360 newsletter. Take a deep breath as there are some really big stories!

Grafana Dazzle and New Relic claim a first!

Arguably, the main Observability event of the past two weeks was ObservabilityCON 2023, where Grafana Labs announced a raft of new releases that take their stack to the next level. Not to be outdone, New Relic have unveiled a private access release of what they are hailing as the industry’s first APM for AI. To complete the trifecta, OpenTelemetry heralded a major performance breakthrough thanks to the power of Apache Arrow. Elsewhere, Victoria Metrics reported some staggering growth figures, whilst a startup from North Carolina are looking to use AI to revolutionise log management.

Feedback

As practitioners in the field, you will know that every good observability system needs a feedback loop. Let us know how we are doing at:

NEWS

Grafana take their stack to the next level

ObservabilityCon 2023 in London got off to a dazzling start as Grafana Labs unveiled a slew of new products and tools. The first major reveal was Application Observability, an application monitoring tool designed to support anomaly detection and help SRE's and developers to reduce MTTR. This was followed by the public launch of Grafana SLO and the news that Asserts.ai are now part of the Grafana family. Arguably the most exciting debut of the day was Beyla, which leverages eBPF to apply autoinstrumentation to a range of languages including Rust and C++.

New Relic open up the AI black box

LLM's represent a whole new world for observability and there was always going to be a certain amount of kudos at stake for being the first company to plant their flag on its soil. It looks like the winners of this particular race are New Relic, who this week announced the release of their AI monitoring product called, naturally enough, AIM. According to the New Relic blog, the product will check for “bias, toxicity, and hallucinations“ as well as identifying processing bottlenecks and scanning for potential vulnerabilities. At the moment the product is in private preview mode, and if you are quick off the mark you can still sign up.

OpenTel add Apache Arrow to their Quiver

Although Grafana and New Relic may have grabbed the spotlight with high profile product releases, one of the biggest stories of the past fortnight was quietly announced on the OpenTel blog, with the unveiling of the OpenTelemetry Protocol with Apache Arrow. This is an impressive breakthrough, with promises to improve compression efficiency by 40% for the majority of workloads. Apache Arrow uses a columnar memory format to achieve massive performance gains for analytics tasks. Interestingly, a major contributor are ServiceNow. Obviously, they are best known for their IT Service Management software, but now appear to be expanding into the observability space.

Victoria Metrics surges ahead

Victoria Metrics has been quietly building up a loyal base of followers over the past few years and has emerged as a viable competitor to established brands such as Prometheus. As well as being a highly capable metrics platform, it is also very competitive on costs. This is clearly a winning combination as they have recorded 320% growth in their Enterprise offering over the past year and have signed up a number of blue-chip customers - including Adidas, Grammarly and Wix. The company was founded in Kiev just five years ago and has now racked up some 268 million product downloads.

LogSail - making waves with AI

LogSail is a startup from Charlotte, North Carolina who are looking to take the toil out of log management. In fact, they are aiming at nothing less than eliminating the need for any more fiddling with YAML or XML files! Their AI-driven product uses intelligent agents to analyse user-defined policies and apply these automatically for forwarding logs from clients to the LogSail API. Last week we chatted to LogSail CEO and founder Stephen Collins, and you can read our article about the company on the Observability 360 web site.

Middleware - a unified Observability Platform

Middleware are a relative newcomer to the observability market. They were founded in September 2022 and include alumni from Splunk, New Relic and Data Bricks among their ranks. They claim to offer a "unified observability platform”. The product provides an all-in-one timeline where you can see logs, traces, and metrics in one place. They have quite an interesting backstory - CEO Laduram Vishnoi apparently founded the company as he was unable to find an “all-inclusive cloud monitoring tool” for his previous startup. If you want to evaluate their system, they offer a Free Forever plan - as well as Pay As You Go and Enterprise tiers.

From the Blogosphere

Mining your logs for buried gold with Mezmo

In recent years, a number of companies have sprung up offering telemetry pipelines to help customers enrich and flexibly route their telemetry data. In this article, Mezmo walk us through a telemetry transformation - showing how pipelines can be used to extract potentially valuable metrics from your logs. In this case they are parsing web logs to derive metrics for describing http traffic.

Building your own AI observability solution

As we mentioned above, New Relic have just released an APM solution for AI applications. You can, of course, roll your own toolkit using readily available observability tools. In this blog article, Bijit Ghosh of Deutsche Bank discusses best practices for observability across the full AI system lifecycle. He composes a custom system which knits together a range of technologies including structlog, Flask, Prometheus and Kibana as well as AI-specific tools such as MLFlow and CausalML. It’s a comprehensive article which exhibits a clear understanding of both observability and AI technologies.

HOW TO

Taking a trip with the Dapr sidecar

Dapr is a framework which is gaining traction amongst software developers and is also one of the building blocks of Radius - Microsoft’s new platform for building cloud-native apps. It uses a sidecar pattern to abstract away concerns such as messaging, logging and persistence - therefore enabling developers to create agile, loosely-coupled services. It also natively supports OpenTelemetry. This article on the Observability 360 web site looks at leveraging the observability capabilities of Dapr to connect to logging, metrics and tracing endpoints.

K8S observability with KubeShark

Kubernetes is a great platform for resilience and scalability. As the number of services deployed increases, so the difficulty of managing network traffic can also grow exponentially. It is not surprising therefore, that there has been a proliferation of tools designed to monitor K8S traffic. In this article, DevOps Engineer Michael Levan kicks off a series of pieces on Kubernetes observability with a quick and easy introduction to KubeShark.

VIDEOS

Unlocking the Kernel - the eBPF story

eBPF has ushered in a Linux revolution, extending the kernel in ways which were previously unthinkable. This documentary features interviews with the engineers who conceived of and built the technology. It is now available to watch for free on YouTube.

Perses - setting standards for visualisation

There are a number of great visualisation tools on the market. At the moment though, there are no standards for templating and deploying dashboards using CI/CD tooling. Perses looks to address this gap with a GitOps-compatible workflow. Prometheus have now made available a recording of the session at this year's Promcon where the project was discussed in detail.

That’s a wrap!

That’s all for this fortnight’s edition. As you may have noticed, the layout of the mailing has undergone a bit of a re-organisation. If you have any feedback on it, please get in touch. We’ll leave you with this quote from the 2021 Splunk State of Observability report:

“With every organization now being a digital organization, observability should be viewed as a core competency, not a cutting-edge differentiator”.