A Comparison of Time Series Databases and Netsil’s Use of Druid

In a previous blog post we talked about the key technologies that we used in developing the Netsil Application Operations Center (AOC), which enables real-time observability and analytics for API and microservices-driven applications. In this blog post, we provide a comparative analysis of various time series databases that we considered for the Netsil AOC and describe how we narrowed down our search to Druid, an up and coming OLAP database.

While legacy monitoring tools rely on analysis of application logs or call stacks, Netsil AOC uses stream processing for capturing live service interactions and decoupling the collection of metrics from the analysis, which is done off the critical path. The time series database is a key component of the Netsil AOC as it enables ad-hoc querying of the metrics within seconds of their ingestion into the data pipeline. The comparison done in this post will be useful for developers, architects and DevOps engineers looking for scalable solutions to retain and query metrics data related to microservices, or other similar use cases.


Druid is a high-performance, column-oriented, distributed data store.


A Python connector for Druid