Customers get complete flexibility on bringing data from their entire stack into Observe using open, vendor-neutral OpenTelemetry Collector, Log Shippers, Prometheus, APIs, and more.

Metrics, Traces, Logs, Events, and business data stored and processed in one unified data lake built on Apache Iceberg – a distributed, community-driven, 100% open-source data table format that enables faster data processing on extremely large datasets.

Observe separates storage from compute so they can be scaled independently, on-demand. This  results in the most efficient utilization of resources. While compute and storage resources are physically separate, they are logically part of a single system that provides elastic, non-disruptive scaling.

Data stored in Observe is compressed between 5 and 50x with metrics (time-series) data compressing much more than logs or traces. Once compressed, all data is stored in low cost object storage.

All data is hot, eliminating the need to create data tiers, migrate data between hot, warm and cold tiers, and deal with the toil associated with rehydrating and reindexing data. Historical data is instantly available for search and analysis.

Compute engines scale up or down elastically to process vast amounts of data quickly and efficiently. Using parallel processing, multiple compute nodes simultaneously operate on the same data to achieve fast query performance – this scales to hundreds of nodes and thousands of cores.

Observe does not require data to be structured or parsed before ingestion – there are no grok scripts. Users can simply point their collector at Observe’s endpoint and benefit from fast, cost efficient ingestion.  Once the data is in Observe’s data lake it can be parsed on-the-fly using schema-on-demand – either visually, through an AI-assisted UI, or using OPAL.

Observe uniquely retains relationships between entities e.g. trace to pod to container so users can ask any question about interdependencies in their distributed system. Users can quickly drill from customers to sessions to services to metrics… and then pivot to logs or trace, all while maintaining context. Observe achieves this by transforming low level event data into structured (and related) materialized views so users can gain actionable insights much faster.

Observe Processing and Analytics Language (OPAL) offers a rich set of features and commands to search, analyze, and transform data at scale.

Built using Generative AI, Large Language Models (LLMs), and Agentic AI, Observe AI assists teams in quickly understanding and learning Observe. O11y Co-pilot can create OPAL queries and regexes instantly using a point-and-click UI, increasing developer productivity. AI Investigator speeds up troubleshooting and reduces time to resolution.