ArkonisDeployment

API: arkonis.dev/v1alpha1 Kind: ArkonisDeployment Short name: aodep

The core resource. Analogous to a Kubernetes Deployment — manages a pool of agent instances running the same model, prompt, and tool configuration. The operator creates and maintains a backing Deployment and Service for each ArkonisDeployment.

Example 

apiVersion: arkonis.dev/v1alpha1
kind: ArkonisDeployment
metadata:
  name: research-agent
  namespace: default
spec:
  replicas: 2
  model: claude-sonnet-4-20250514
  systemPrompt: |
    You are a research agent. Gather and summarize information
    accurately and concisely. Always cite your sources.
  mcpServers:
    - name: web-search
      url: https://search.mcp.internal/sse
    - name: memory
      url: https://memory.mcp.internal/sse
  limits:
    maxTokensPerCall: 8000
    maxConcurrentTasks: 5
    timeoutSeconds: 120
  livenessProbe:
    type: semantic
    intervalSeconds: 60

Spec fields

Top-level

Field Type Required Description
replicas int32 no Number of agent pod replicas. Defaults to 1.
model string yes Model identifier passed to the configured LLM provider (e.g. claude-sonnet-4-20250514 for Anthropic).
systemPrompt string yes System prompt injected into every API call made by agent pods.
mcpServers []MCPServerSpec no List of MCP servers to connect at pod startup. See MCP Servers.
limits AgentLimits no Per-agent resource and token limits.
livenessProbe AgentProbe no Semantic health check configuration.
configRef string no Name of an ArkonisConfig in the same namespace. Merged into effective system prompt and model settings.
memoryRef LocalObjectReference no Name of an ArkonisMemory in the same namespace. Injects memory backend config into agent pods. See Agent Memory.

mcpServers[]

Field Type Required Description
name string yes Logical name. Used as tool name prefix (name__tool).
url string yes SSE endpoint URL of the MCP server.

limits

Field Type Default Description
maxTokensPerCall int 8000 Token budget (input + output) per LLM API call.
maxConcurrentTasks int 5 Max tasks a single pod processes simultaneously.
timeoutSeconds int 120 Per-task timeout. Task is abandoned and an error is returned after this duration.

livenessProbe

Field Type Default Description
type string ping ping — HTTP reachability only. semantic — enables /readyz API validation.
intervalSeconds int 60 Interval between semantic validation checks.
validatorPrompt string (internal default) Custom prompt sent to the LLM during /readyz semantic validation. Falls back to a built-in default if not set.

Status fields

Field Type Description
replicas int32 Total number of agent pods managed by this deployment.
readyReplicas int32 Number of pods passing both liveness and readiness checks.
conditions []Condition Standard Kubernetes conditions (Available, Progressing, Degraded). 
kubectl describe aodep research-agent
# ...
# Status:
#   Ready Replicas:  2
#   Replicas:        2
#   Conditions:
#     Type:    Available
#     Status:  True

Apache 2.0 · ARKONIS