Overview
Exoscale Managed Thanos delivers a fully managed, high-availability deployment of the popular Prometheus long-term storage and global query solution. The service simplifies the management of large-scale metrics systems, allowing teams to focus on deriving insights from their data rather than operating complex infrastructure. It is designed for organizations that need scalable, cost-effective metric storage and querying across multiple Prometheus instances.
Terminology
- Service a running Thanos cluster managed by Exoscale DBaaS
- Plan a sizing tier based on the number of Prometheus clusters to connect
- Query component that aggregates and queries metrics from both real-time receivers and historical storage
- Query Frontend caches query results and splits large queries into smaller sub-queries for efficient execution
- Receiver handles metrics ingestion via Prometheus remote write, collecting data in real-time
- Store Gateway interfaces with object storage to provide access to historical metrics data
- Compactor downsamples and compacts metrics in object storage to optimize storage usage and query performance
Architecture
Thanos combines multiple components into a unified cluster architecture with seamless data flow:
Data Collection & Storage Receivers ingest metrics in real-time via Prometheus remote write. After each Time Series Database (TSDB) block completes (typically every 2 hours), data is uploaded to object storage for long-term retention.
Query Processing The Query Frontend receives requests and forwards them to Query components, which retrieve real-time data from Receivers or historical data from the Store Gateway. Duplicate samples from multiple receivers are automatically removed to maintain data integrity. Query results are cached by the Query Frontend to accelerate subsequent requests.
Storage Optimization The Compactor runs in the background to downsample and compact historical data, reducing storage costs and improving query performance for older metrics.
Features
- Managed Cluster Lifecycle automated provisioning, patching and minor-version upgrades
- Unlimited Metric Retention cost-effective long-term storage using object storage with automatic compaction
- Global Query View query metrics across multiple Prometheus instances through a unified interface
- Prometheus Compatibility fully compatible with Prometheus and standard tooling like Grafana
- Automatic Downsampling reduces storage needs and improves query performance for historical data
- Query Caching Query Frontend caches results to accelerate repeated and similar queries
- High Availability deduplication across receivers ensures consistent data despite multiple collection points
- High SLA 99.95 % on Startup, 99.99 % on Business plans
- Portal, CLI & API Access manage services via Exoscale Portal, REST API and CLI
Plans
Startup Plans (Single Node)
Best for development, testing, or smaller production workloads. SLA: 99.95%
| Plan | vCPU | RAM | Storage | Price |
|---|---|---|---|---|
startup-4 | 2 | 4 GB | 160 GB | See pricing |
startup-8 | 4 | 8 GB | 320 GB | See pricing |
startup-16 | 4 | 16 GB | 640 GB | See pricing |
startup-32 | 8 | 32 GB | 1.28 TB | See pricing |
startup-64 | 12 | 64 GB | 2.56 TB | See pricing |
Business Plans (3-Node HA Cluster)
For production workloads requiring high availability. SLA: 99.99%
| Plan | vCPU/node | RAM/node | Storage (total) | Price |
|---|---|---|---|---|
business-4 | 2 | 4 GB | 480 GB | See pricing |
business-8 | 4 | 8 GB | 960 GB | See pricing |
business-16 | 4 | 16 GB | 1.92 TB | See pricing |
business-32 | 8 | 32 GB | 3.84 TB | See pricing |
business-64 | 12 | 64 GB | 7.68 TB | See pricing |
For detailed configuration options, see the Reference section.
Availability
Thanos is available in all current Exoscale zones.
| Zone | Country | City | Availability |
|---|---|---|---|
at-vie-1 | Austria | Vienna | |
at-vie-2 | Austria | Vienna | |
ch-gva-2 | Switzerland | Geneva | |
ch-dk-2 | Switzerland | Zurich | |
de-fra-1 | Germany | Frankfurt | |
de-muc-1 | Germany | Munich | |
bg-sof-1 | Bulgaria | Sofia | |
hr-zag-1 | Croatia | Zagreb |
Limitations
- No Terraform Provider Exoscale Terraform provider doesn’t support Thanos yet
- No External Thanos Access cannot connect Exoscale-managed services to external (non-Exoscale) Thanos endpoints; must use Exoscale Managed Thanos for integrations
- Single-Zone Clusters services run within a single zone; cross-zone replication is not currently available
Quick Links
Exoscale Resources
- Exoscale Portal - Create and manage Thanos services
- REST API Documentation - Full API reference
- Pricing - Current plan pricing
Exoscale DBaaS General
Thanos Service
- Retention Rules & Downsampling - Data retention and downsampling configuration
External Resources
- Thanos Project - Official Thanos documentation
- Aiven for Metrics - Underlying platform documentation
- PromQL - Query language reference