Navigating the Growth: Choosing Between Thanos and Mimir for AKS Clusters


Our company is on the cusp of a major expansion, planning to scale our infrastructure to encompass as many as 500 AKS clusters. In light of this growth, we’re in the process of assessing enduring monitoring solutions to guarantee consistent oversight and dependability throughout our network. Our main options under consideration are Thanos and Mimir.

These two candidates appear to provide comprehensive features for augmenting Prometheus-based monitoring systems, including sustained storage and efficient management of extensive metric data. Nonetheless, considering the breadth of our deployment, we aim to choose a solution that is validated not only by its technical merits but also by practical, hands-on experiences.


is recognized for its straightforward integration with current Prometheus frameworks and its economical approach to long-term storage through object storage. Its capabilities for a unified query perspective and deduplication across clusters are especially attractive.


, on the other hand, is a more recent extension of Prometheus that boasts enhanced scalability and performance tweaks tailored for expansive operations. Its design and query functions are optimized for high efficiency in large-scale settings.

We seek the community’s input on several critical aspects:


: Can Thanos and Mimir effectively scale alongside hundreds of clusters? We’re keen on understanding both the operational impact and the resource utilization efficiency.


: We would value any feedback on the dependability of these systems when deployed extensively. How are failures managed, and what does the recovery entail?


: What are the query response times and data retrieval speeds like for these solutions? Insights into their performance, particularly for real-time monitoring, would be highly beneficial.


: Although not our primary focus, we’re interested in the long-term financial considerations of implementing either system, especially in terms of storage and computational demands.

Ease of Use and Integration

: How user-friendly are these solutions when integrating with AKS, from initial setup to everyday management?

For those who have experience with Thanos or Mimir at a comparable scale, your shared experiences, obstacles encountered, lessons learned, and any advice would be immensely appreciated.

We thank you in advance for your valuable insights.


Thanos is lauded for its ease of integration with existing Prometheus setups. It utilizes object storage for cost-effective long-term data retention and offers a global query view, which is crucial for maintaining visibility across numerous clusters. Its deduplication feature ensures that data is not unnecessarily replicated, saving on storage costs and simplifying data management. Mimir, although newer, is designed with large-scale deployments in mind. It promises high scalability and performance optimizations, which are essential when dealing with hundreds of clusters. Mimir’s architecture is built for efficiency, aiming to reduce the operational overhead and resource consumption that typically come with scaling up.


Both Thanos and Mimir are designed to scale with Prometheus. Thanos achieves this through a modular approach, allowing you to add components as needed to handle additional load. Mimir, leveraging its microservices-based architecture, can also scale out to accommodate growth in infrastructure.


In terms of reliability, both systems are built to handle failures gracefully. Thanos and Mimir support high availability configurations, ensuring that monitoring continues even if a component fails. Recovery processes are streamlined, minimizing downtime and maintaining continuous monitoring.


Performance is a key factor, especially when it comes to query latency and data retrieval times. Both solutions are optimized for high-load environments, with Thanos providing a more traditional approach to data storage and retrieval, while Mimir focuses on performance at scale, potentially offering faster query responses due to its architectural optimizations.


While cost is not the primary concern for many organizations, it is still a consideration. Thanos, with its object storage integration, can be more cost-effective for long-term data retention. Mimir’s efficiency at scale may translate to lower operational costs as the infrastructure grows.

Ease of Use and Integration

Ease of use and integration with AKS are crucial for daily operations. Thanos is known for its simplicity and minimal operational overhead. Mimir, being newer, may require a steeper learning curve but is designed to integrate seamlessly with modern cloud-native environments.

In conclusion, both Thanos and Mimir offer compelling features for scaling Prometheus-based monitoring. The choice between them may come down to specific organizational needs and preferences. Those with existing Prometheus setups may find Thanos to be a natural extension, while organizations looking for cutting-edge performance optimizations might lean towards Mimir. Ultimately, real-world experiences and community feedback will be invaluable in making an informed decision for such a critical aspect of infrastructure management.

Leave a Reply

Your email address will not be published. Required fields are marked *

Privacy Terms Contacts About Us