[Podcast] SaaS Observability Done Right

When we talk about next-gen SaaS products, Product-Led Growth (PLG) and bottom-up terms come to mind. What are these all about? What do developers need to look out for? Aviad Mizrachi, co-founder and CTO at Frontegg and Dotan Horovits, Technology Evangelist at, converged to discuss the crucial SaaS Observability aspect. Here’s a quick recap.

What is SaaS Observability?

SaaS Observability helps companies understand how an app is operating based on the data that is harvested from end-user and system interactions. This data can come from product usage patterns or metrics generated via the infrastructure. Observability should not be confused with Controllability, which is the ability of external factors to impact performance and other related KPIs. 

When it comes to SaaS, Observability requires careful architectural design planning. As you’ll learn in this podcast, both frontend and backend aspects need to be covered with a well-orchestrated toolbox of third-party solutions.

Aviad Mizrachi: “SaaS Observability is Key Today”

The shift towards PLG means that the entire focus is on product features and how they are performing at any given time. With Frontegg, was one of the first solutions that was implemented, first on a trial basis. Once the metrics started flowing in and the value was established, this became a match made in heaven. The company gained the ability to know what customers are doing. 

“The experience is now expected to be spotless,” Mizrahi explained.

“If your product is providing APIs, the entire movement goes to the backend. You have to make sure that the observability should be back-end centric, with API-call observability and aggregated views becoming key aspects. We started using Heap and MixPanel for user analytics. Frontegg has also implemented FullStory and SmartLook to record sessions and recognize friction points.“

Dotan Horovits: “Observability has Multiple Stakeholders”

SaaS observability goes beyond the developers, it ropes in multiple stakeholders. This is challenging since optimal cross-department collaboration is required on an ongoing basis. You have marketing teams that need to track leads and their status, product experts that need to act upon insights, and also the development teams that need to add or tweak features.

“ believes that if there is data, it must be collected,” Horovits elaborated.

“The initial onboarding is just the beginning,“ Horovits continued. “Every new feature is like a mini-release for companies. There are new KPIs and new dashboards for each and every feature. With, we have a bunch of professionals, including business intelligence experts, breaking down the new features and syncing the new dashboards with the product’s master-KPIs.“ A Key Part of Frontegg’s Developer Stack helps you search and analyze logs, metrics and traces to understand the state of your system at any given time and perform root-cause analysis on issues as soon as possible, something that’s extremely crucial in today’s competitive market. With the company’s high-performance filters, you get the ability to examine your logs without using any type of query language. 

This basically lets you accelerate troubleshooting with the help of actionable insights gained via checking log patterns. Getting started is easy. You just need to implement a few lines of code to get going. also gives you full observability into your active environments via a centralized and user-friendly dashboard. This gives you the ability to track all relevant traces, metrics and logs in real-time. You can also get notified about all kinds of events and anomalies by customizing your triggers as per your needs and requirements. Alerts can be sent to PagerDuty or Slack and also via email.

The Bottom Line

Aviad Mizrachi wrapped up proceedings by stressing the importance of separating hot data from cold data, which is essentially anything older than three months. That being said, cold data doesn’t necessarily need to be deleted and can be stored in any third-party storage. This cold data can play a role when conducting annual analysis and breaking down long-term feature performance.

“Collect as much data as you can,” Mizrachi concluded.

Product-Led Growth is all about collecting data and crunching it as soon as possible to make an impact and grow your business. When it comes to tooling, he recommended adopting market-leading solutions and finding the right blend. The solutions need to be easy to use for remote staff or new workers, while having high visibility for optimal cross-team collaboration. It’s time to get PLG-d.