Skip to content

AxonOps — AI-Native Control Plane for Open Source Data Platforms

System requirements

Minimum system requirements for installing AxonOps Server and AxonOps Dashboard on a self-hosted server to monitor and manage the 3 nodes that are part of the Starter Tier.

This is not a system requirements spec for Cassandra nodes where the AxonOps Agent is installed.

ComponentRequirement
Processor1 gigahertz (GHz) or faster with 4 or more cores on a compatible 64-bit processor
RAM8 gigabytes (GB)
Storage20 GB or larger storage device
ComponentRequirement
Processor1 gigahertz (GHz) or faster with 4, ideally 8 or more cores on a compatible 64-bit processor
RAM16 gigabytes (GB)
Storage120 GB or larger storage device

If you need to increase the number of nodes to be monitored, get in touch with us.

The AxonOps Dashboard is a web application and must be accessed with a browser built on the Chromium engine. Browsers built on other engines, such as Mozilla Firefox (Gecko) and Apple Safari (WebKit), are not supported. On an unsupported browser the Dashboard may exhibit unexpected behaviour: parts of the interface may render incorrectly, fail to load, or not respond as intended.

BrowserEngineSupported
Google ChromeChromium
Microsoft EdgeChromium
BraveChromium
OperaChromium
VivaldiChromium
Mozilla FirefoxGecko
Apple SafariWebKit

Use a current release of the chosen browser. Chromium-based browsers not listed above are expected to work but are not tested.

A single axon-server instance is capable of monitoring well over 1,000 nodes across multiple clusters. axon-server is not designed for clustered deployment — horizontal scaling is handled by the underlying datastores (Elasticsearch or OpenSearch for logs, events, and configuration; optionally Cassandra for metrics at larger deployments).

For deployments above 8 monitored nodes, using Cassandra as the metrics store is recommended. See Using Cassandra as a Metrics Store for configuration details.

The following formulas apply to the axon-server process itself. Datastores (OpenSearch, Elasticsearch, or Cassandra for metrics) are sized separately.

Let n = total number of monitored nodes.

ResourceFormula
CPU limit (milliCPUs)500 + (n × 50)
Memory limit (MiB)1024 + (n × 128)

Example values:

Monitored nodesCPU limitMemory limit
201500m3584 MiB (~3.5 GiB)
503000m7424 MiB (~7.25 GiB)
1508000m20224 MiB (~19.75 GiB)
30015500m39424 MiB (~38.5 GiB)

Set resource requests to 50% of the computed limits as a starting point and adjust based on observed utilisation.

Apply the same formula to size the host running axon-server, rounding up to the nearest even number of vCPUs:

ResourceFormula
vCPUround up to nearest even: (500 + (n × 50)) / 1000
RAM (GiB)round up to nearest even: (1024 + (n × 128)) / 1024

Example values:

Monitored nodesvCPURAM
2024 GiB
5048 GiB
150820 GiB
3001640 GiB

These figures cover the axon-server process only. Allow additional headroom for the operating system and any co-located components (e.g. axon-dash).