Observability has a data problem. AI is about to make it worse.

Observability is facing an emerging challenge as artificial intelligence systems generate increasingly large volumes of logs, traces, and metrics. This exponential growth in telemetry data is exacerbating the industry's longstanding data storage problem, making it difficult for teams to gain full visibility into their systems. Dublin-based Bronto, an intelligent data observability platform, believes the next battleground in observability will be at the data layer, rather than the dashboard layer.

Trevor Parsons, co-founder and co-CEO of Bronto, argues that the industry has been optimizing at the edges rather than addressing the core issue of telemetry storage. He notes that the industry has introduced numerous "hacks" and "capabilities" to avoid tackling this issue head-on and protect their margins. Parsons emphasizes that the cost of storing data, especially in the AI era, is becoming unsustainable for many teams, with some paying up to 20-30% of their infrastructure spend on observability.

Observability has always struggled with a data problem, but in practice, the view teams get is often incomplete, expensive, and short-lived. Teams are forced to compromise on retention, sampling, and data access due to the limitations of existing tools, which ultimately leads to blind spots in observability. The solutions provided by vendors often come with compromises, forcing teams to choose between cost, coverage, and speed of data access.

The traditional observability business model charges customers based on data storage, regardless of how much value they derive from the data. This model doesn't align with customer value, as customers often find existing tools difficult to use and feel that their observability solution is just a costly data store. Noel Ruane, co-founder and co-CEO of Bronto, suggests that teams should pay less for idle data and more when they actually derive value from it through queries and analysis.

Bronto has developed a custom-built polymorphic data store called BrontoD, designed specifically for observability data. This platform allows enterprises to retain more than 100 times the observability data they currently hold, without compromising performance or usability. The company's technical differentiation lies in its focus on addressing the data problem before tackling the visualization aspect of observability.