The new version of ESM is here!

ESM, our Observability platform for Data Scientists, has reached version 2020.1. It is our biggest release to date.

Nikola Markovic

21 May 2020 · 4 mins

This week we’re releasing our biggest ever update to ESM. Here’s a summary of my favourite new features.

A whole new Experience

Back in 2012, those MSOffice-inspired ribbons were still very popular, so ESM v1 ended up with a ‘toolstrip and tabs’ style user interface. As we added more functionality, accessing everything through that single toolstrip started to become a little confusing, and over time, some new users even started telling us that they found the sheer number of buttons intimidating.

It’s like having a Ferrari that I don’t know how to drive!

A. Customer

So as of this 2020 release, we’ve completely rebuilt the ESM interface. The UI has been completely reorganised, we’ve implemented a few usability features, and we’ve given it a bit of a style makeover. We think it looks a lot cleaner. Check out the before and after.

ESM also now fully supports internationalization (i18n). If you’re an international customer and you’re willing to help us validate our translations to your language, please get in touch!

First class support for SAS® Viya

We spent much of last year working closely with SAS R&D in Cary to make sure that ESM can fully support monitoring SAS Viya. As a result, Viya is now probably our best-supported platform, and we’ll be releasing more features specific to it in the coming months (that graph a couple of paragraphs down is a small preview of that).

It was also the work we did with SAS R&D that ultimately drove many of the new features and integrations mentioned in this post. I’m personally especially grateful to Craig Rubendall and Paul Kent for the time they invested in helping us make our product better. It was a real pleasure working with you guys!

Spans, advanced annotations

ESM has always supported tags - point events that enable features like step annotation of user code. With this release of ESM, we’ve also added basic support for spans. The CAS action annotations explaining what this CAS session is doing are a great example of spans.

Span support is also about to get a lot more interesting in our next release, with the changes we’re introducing to our data model. Watch this space!

Efficiency improvements, Container support, workload Autodiscovery

Under the covers, we have made a number of improvements to our agent architecture. Agents can now be configured to autodiscover and auto-label user activity, which simplifies many integrations and lets us identify ‘rogue’ workloads when performing cloud migration sizing. They can also now monitor and identify Kubernetes workloads from inside a container. More Kubernetes-specific features are coming in the next few months.

My favourite thing about our latest agent though, is the CPU footprint. It can now autodiscover, collect, serialise and send 2,000 processes worth of data back to the ESM server in real time, and use less system resource than a single instance of htop. Around 4-5% of a single CPU thread. And that’s just awesome.

Closing thoughts

Chris, the one in charge of making this release happen, speaks with great pride about the 485 merges, 160 issues, and other statistics related to it. And quite rightly so - those numbers are huge, and Chris also deserves a medal for the way he’s dealt with my constant attempts at scope creep.

But there’s a little more to it than that. Our development team at Boemska has more than doubled in size over the last 12 months, and to me this milestone means that much more because it is proof of our capability to deliver while we scale as a software company - something I always knew would be pretty hard to do. So I’m extremely proud of everyone who made this happen. Looks like we have an exciting year ahead of us.