Skip to main content

Data Tiering

Levitate supports automatic data tiering of metrics based on retention policies and query range limits which allows long term retention of time series data without any downsampling or rollups.

Levitate supports automatic data tiering of the metrics based on retention policies. Leveraging this underlying storage capability, Levitate can store data for long-term durations without downsampling or rollups. It enables traffic shaping of time series data by creating access policies tailored for specific use cases. It is a fundamental differentiator, making Levitate a time series data warehouse built for long-term retention and superior performance.

Each Data Tier has a retention period, which dictates how long it will retain the data, and a query range, which dictates how far in the past one can query.

Additionally, its tier has a query depth, which decides how much one query is in a single query.

Blaze Tier​

Ideal for Alert Manager, real-time, shallow consumption of data. Retains data for two hours. It has delta caching enabled for faster queries.

Hot Tier​

Ideal for quick daily trends and impact history of recent changes and fixes. Retains data for six months. Best for exploratory workloads such as Grafana dashboards.

Cold Tier​

Ideal for long-term data exploration and analysis. Levitate offers long-term retention of a year of data retention without any rollups or downsampling out of the box. This is customizable as per customers' requirements and reliability requirements and can be extended even further.

Levitate Cluster: Data Tiers

You can access the tiers from a Levitate cluster's data tiers tab.

How does it work under the hood?​

Levitate automatically tiers the data. Users do not have to do anything special. You can follow the instructions in the Levitate onboarding guide and start sending metrics, which will be automatically tiered based on the retention policies. After that, you can create read tokens for specific tiers from the Policies section.

More about reading data from different tiers can be found in the Read data from Levitate guide.