blueflood

If you're testing those, it would be nice if you could test and make a comparison with the cassandra-based Blueflood as well. https://github.com/rackerlabs/blueflood/wiki

https://github.com/rackerlabs/blueflood is a time series datastore with aggregation and rest APIs built on top of Cassandra. It's production ready and in fact used in production at Rackspace. Full disclaimer: I am a core contributor.

Rackspace, Inc. - San Francisco, CA - Software Engineer ------------------------ ------ Who We Are ------ ------------------------ We are the Cloud Metrics team at Rackspace. We run a metrics-as-a-service API. We created the Blueflood open-source metrics engine and API (http://blueflood.io). We make it easy for developers to store and scale their metrics. ------------------------- --- Software Engineer --- ------------------------- Focus on time-series data and make Cloud Metrics a world-class engine for metrics of all shapes and sizes. Scale Cloud Metrics' infrastructure: help our large and growing Cassandra cluster run as smoothly as possible. Engage the growing open-source community around time-series data and metrics analysis and visualization. Optimize Cloud Metrics' deployment, continuous integration and testing proces

Rackspace, Inc. - San Francisco, CA - Software Engineer ------------------------ ------ Who We Are ------ ------------------------ We are the Cloud Metrics team at Rackspace. We run a metrics-as-a-service API. We created the Blueflood open-source metrics engine and API (http://blueflood.io). We make it easy for developers to store and scale their metrics. ------------------------- --- Software Engineer --- ------------------------- Focus on time-series data and make Cloud Metrics a world-class engine for metrics of all shapes and sizes. Scale Cloud Metrics' infrastructure: help our large and growing Cassandra cluster run as smoothly as possible. Engage the growing open-source community around time-series data and metrics analysis and visualization. Optimize Cloud Metrics' deployment, continuous integration and testing proces

Rackspace, Inc. - San Francisco, CA - Software Engineer ------------------------ ------ Who We Are ------ ------------------------ We are the Cloud Metrics team at Rackspace. We run a metrics-as-a-service API. We created the Blueflood open-source metrics engine and API (http://blueflood.io). We make it easy for developers to store and scale their metrics. ------------------------- --- Software Engineer --- ------------------------- Focus on time-series data and make Cloud Metrics a world-class engine for metrics of all shapes and sizes. Scale Cloud Metrics' infrastructure: help our large and growing Cassandra cluster run as smoothly as possible. Engage the growing open-source community around time-series data and metrics analysis and visualization. Optimize Cloud Metrics' deployment, continuous integration and testing proces

Rackspace, Inc. - San Francisco, CA - Software Engineer ------------------------ ------ Who We Are ------ ------------------------ We are the Cloud Metrics team at Rackspace. We run a metrics-as-a-service API. We created the Blueflood open-source metrics engine and API (http://blueflood.io). We make it easy for developers to store and scale their metrics. ------------------------- --- Software Engineer --- ------------------------- Focus on time-series data and make Cloud Metrics a world-class engine for metrics of all shapes and sizes. Scale Cloud Metrics' infrastructure: help our large and growing Cassandra cluster run as smoothly as possible. Engage the growing open-source community around time-series data and metrics analysis and visualization. Optimize Cloud Metrics' deployment, continuous integration and testing proces

Cassandra is very good at storing ordered sets of data which you can then pull out the nearest value to a key. There are a load of layers on top of this that are more specifically time-series which it probably makes sense to use rather than building your own: - https://code.google.com/p/kairosdb/ - https://github.com/OpenNMS/newts - https://github.com/rackerlabs/blueflood - https://github.com/pyr/cyanite

I think you have most of it right. The rollups are not configurable currently. However, we did give some thought into those pre-configured rollup granularities. The query today is very simple. You can query data points for a metric you know by name. Look at retrieve metrics section of this guide: https://github.com/rackerlabs/blueflood/wiki/10minuteguide I'd like to clarify that this is not a monitoring tool. This is a metrics aggregation engine for data collected using monitoring tools. We haven't benchmarked OpenTSDB. We have a lot of operational experience with Cassandra and one of the devs is a Cassandra committer. We are seeing good performance in terms of reads and writes from/to Cassandra. If you have other questions, please join #blueflood on freenode. I am one of the devs and I am in that channel too.