Runtime limits
The runtime limits are account-wide: how many times a loop may iterate, how many rows a database call may return, and how long an agent definition may be. Developers edit them in the UI under Guardrails, and they apply to every agent in the account. How long a run may take and how much it may spend on LLM calls are set for each agent; see Time and cost bounds.
How the limits are checked
The limits are checked while the agent runs. A long loop yields at the interval that the yield loops after setting gives, so that the agent's time bound applies inside the loop. A loop that exceeds the iteration limit fails the run. A db command returns at most the row limit. The definition length applies to the JSONC text of a draft, comments included. The stream heartbeat sets how often a streamed response sends a keep-alive line.
| Field | Type | Default | Description |
|---|---|---|---|
maxIterationLoops | integer | 10000 | The most iterations a 'while' or 'for' may run. A loop inside a loop is counted separately. |
yieldLoopsAfter | integer | 1000 | How many iterations a loop runs before yielding, so the runtime limit can be checked. |
dbMaxRows | integer | 10000 | The most rows a single db command may return. |
agentMaxLength | integer | 20480 | The most characters an agent definition may contain. |
streamHeartbeat | integer | 15 | How often a noop is sent on an idle output stream, in seconds, so browsers and proxies do not close it while the agent works. |

