I recently had the opportunity to attend a technology briefing from Uila, a company that works in the IT Monitoring space. This is a briefing I took up with great interest since colleagues and friends in the community had been talking (and writing) highly of them after they had presented at Tech Field Day 13 (links to videos and written articles, look up through the page for them). Here’s a bit of what they do and some of my thoughts about it.
Uila is a startup launched in 2013 in Santa Clara, Silicon Valley. The three co-founders (Chia-Chee Kuan, Dean Au and Miles Wu) have all extensive industry background in the networking area, and it’s no wonder that this is strongly felt in their product, Uila, which also bears the same name as the company.
The Uila team have built an application-centric, agentless monitoring solution that relies on deep network packet analysis. This allows the tool not only to show the status of an environment from a full stack perspective (from the application/component layer all the way down to the infrastructure), but also to understand the interactions between applications and services if the traffic is not encrypted, and to do so up to the transactional level. This is awesome, although also a bit terrifying at the same time.
The interface shows several layers (services, applications, VM, infrastructure) and can be “peeled” like an onion to deep dive into the elements. Several graphs and views provide a drill-down of the infrastructure and by now you’ll have guessed that the UI component is critical to Uila’s success. They have a fantastic interactive demo available on their website to get an idea. A good little plus is that the interface can be customized in terms of style/color schemes in order to accommodate users who might have difficulties with dark themes, or on the opposite users who can’t suffer light themes.
Beyond visualization, collection of metrics and dependency mapping, Uila can also use this extensive knowledge to propose a probable root cause analysis of the problem thanks to machine learning, although I had no chance to see this in action. There is more than just this, there are logs, and also a kind of recording of past monitoring events, although I do not know exactly if this happens automatically or if users have to record them somehow (I’d bet on automated recording though).
On a final note, Uila supports also multi-tenancy. Organizations which provide infrastructure services to different internal stakeholders or child companies may find this option interesting, as a good way to present their customers with an partial view of the infrastructure. The same may also apply for service providers.
The ability of analysing data comprehensively from multiple sources and aggregating them is obviously critical to get a clear understanding of what’s happening in one’s environment and pinpoint the probable root cause of an issue. Several monitoring vendors do this, and do it well.
What I like undoubtedly with Uila’s approach is the technique employed to do so, via deep network packet analysis, which is tremendous as you may have read above. Obviously, I’m not a networking guy so I find this to be absolutely amazing, especially the ability to get down up to the transactional level and understand what kind of communication happens between services. This could be helpful to determine issues that are not related to the infrastructure, such as to how a DB query works (or not), for example.
While I do respect the considerable engineering effort put into the product architecture, I also have a weak spot fo UIs and appreciate the almost obsessive amount of work and thought that must have been put into designing a beautiful, comprehensive and intuitive user interface. Users can drill down into almost every component or metric, so the most curious/inquisitive minds may get bang for their buck.
The fine folks at Uila have created a very interesting product that delivers outstanding business outcomes, while keeping true to their network engineering roots. They managed to maintain a solid data collection and correlation background, while presenting the findings is a totally amazing way, which allows users to delve and revel into an intense, visually rich IT monitoring experience.