Google models GKE costs for agent sessions: parking idle sandboxes with Agent Substrate saves 40% of the node bill for 100 coding agents at 2 prompts an hour
An agent session stays open all day and mostly waits for a person, so scaling it like web requests pays for the waiting. A cost model and its break-even points, for teams moving coding agents onto a shared cluster.
// Key points
- Citing a study of GitHub Copilot agent traces, the agent waits for the user about 80% of the time in a typical multi-prompt session, and a turn averages 396.3 seconds of agent work. Assuming 2 prompts an hour over an 8-hour day, the agent works 22% of the time.
- With one Pod per session, 100 coding agents need 34 nodes and about $1,288 a month. Snapshotting idle actors with Agent Substrate after a 2-minute timeout needs a 15-node warm pool plus one control-plane node, saving 40% of the node bill; the pool is full about once in 3,500 turns, when a request waits 5 seconds by default and then gets HTTP 503.
- Busier agents save less: at 4 prompts an hour the pool needs 27 nodes and saves 4%, and above about 4.5 one Pod per session is cheaper. The control plane is a fixed cost, so parking pays from about 37 coding sessions. Overcommitting Pods at a 0.25 vCPU request with a 1 vCPU limit fits the same 100 agents on 9 nodes, at the price of agents competing for CPU.
- Agent Substrate does not suspend idle actors by itself: your harness must call the suspend API after a timeout. Autoscaling follows worker Pods, not suspended actors, so scale the pool to zero after hours with a scheduled job.
Builder's takeRead with the 11-11 item on NVIDIA donating Agent Substrate to the CNCF, this is the first cost model you can apply directly: measure your prompt rate before choosing. Sessions under 4 prompts an hour suit parking; busier ones should get one Pod each or be overcommitted. PandaClaws' background jobs mostly fire intermittently, so I would instrument how busy its sessions really are first.
// Background · from #AI agents
Full timeline →- Nov 17 Amazon WorkSpaces adds agent governance: OIDC sign-in, on-behalf-of sign-in as the user, and session replay stored in your own S3
- Nov 17 Google Threat Intelligence: attackers now use autonomous LLM frameworks to scan software updates, work out the patched flaw and build an attack without human involvement
- Nov 16 Anthropic breaks down the architecture of always-on agents into five primitives: event inbox, timers, memory, permissions and outbox