AI Agent Operations

A noisy tenant cannot starve everyone else,
fairness queues per agent, per tenant

Agent Burst is the per-agent-type bounded queue. Each tenant gets a per-tenant share of each agent's capacity. During a tenant's incident, that tenant's tasks burst above the share, but never to the point of starving other tenants. The platform stays responsive even during bad days.

Get Started Talk to Sales
app.novaaiops.com / agent-burst
● LIVE
Per-agent
capacity caps
Per-tenant
share floors
Burst
allowed, not starvation
Live
queue depth visible
How Bursting Works

Share floor, capacity ceiling, fair drain

Each tenant gets a guaranteed share (default = capacity / number of active tenants). When one tenant has spare capacity from its share, that capacity is offered to bursting tenants in FIFO order. When a tenant's share is needed back, in-flight tasks finish (no kill) and queued tasks for that tenant move to the front. No tenant ever starves; no tenant ever monopolizes.

  • Guaranteed share: every tenant always has its share floor available; cannot be displaced by bursters
  • Burst into spare: unused share is offered to bursters FIFO; first come, first served
  • Reclaim is gentle: returning to share never kills a task; in-flight finishes, then capacity returns
app.novaaiops.com / agent-burst · math
Per-Agent Tuning

Different agents, different limits

Each agent type has its own capacity. Cheap, fast agents (log-triager on Haiku) might have capacity 32. Expensive, slow agents (incident-summarizer on Opus) might have capacity 4. The page shows current capacity, current usage, and a recommendation when the queue is consistently bottlenecked on one agent type.

  • Per-agent capacity: tuned to model cost, model latency, and concurrency budget
  • Live usage: concurrent count and queue depth per agent type, refreshed every 5s
  • Bottleneck recommendations: when one agent's queue is consistently full, the page suggests increasing capacity or routing to a cheaper variant
app.novaaiops.com / agent-burst · per-agent
Tenant Reporting

Every tenant sees its own utilization

Each tenant can see its own utilization on the page: how often you bursted past your share, how often you queued, how often a queued task timed out. The data helps a tenant make capacity-vs-cost decisions for their own usage. Other tenants' usage is never visible, only your own.

  • Per-tenant view: shows your share, your bursts, your queue waits, your data only
  • Capacity decisions: use the data to decide whether to upgrade your capacity tier
  • No cross-tenant leakage: tenants only see their own numbers; aggregate across tenants is platform-admin only
app.novaaiops.com / agent-burst · tenant
Pause and Drain

Tenant-side maintenance, gracefully

When a tenant wants to pause its agent activity (deploy freeze, maintenance, security review), one click drains its share gracefully: in-flight tasks finish, no new tasks dispatch. Other tenants are unaffected. Resume is also one click.

  • Graceful drain: in-flight tasks complete normally; new dispatches halt
  • No cross-tenant impact: other tenants keep working; their share is never affected by your pause
  • Time-boxed: pauses are time-boxed by default (24h) so you do not forget to resume
app.novaaiops.com / agent-burst · pause
Video walkthrough coming soon

Subscribe to Nova AI Ops on YouTube for demos, tutorials, and feature deep-dives.

Multi-tenant fairness without per-tenant pain

Bursting is fine. Bursting that locks out another tenant's incident is not. The queue makes the difference automatic.

Get Started Request a Demo