Interface representing a performance record for an individual client within a process. Captures detailed execution metrics, memory, and state for a specific client (e.g., a session or agent instance), used to analyze performance at the client level. Embedded within IPerformanceRecord.clients to provide per-client breakdowns, likely logged via ILogger or emitted via IBus for monitoring (e.g., in ClientAgent workflows).
clientId: string
The unique identifier of the client, typically a session or agent-specific ID. Matches the clientId used in runtime params (e.g., this.params.clientId in ClientAgent), linking performance data to a specific session or agent instance. Example: "client-456" for a user session.
sessionMemory: Record<string, unknown>
A key-value record of the client’s session memory.
Stores arbitrary data (e.g., cached values, temporary variables) used during the client’s operation, similar to IState’s state management in ClientAgent.
Example: { "cacheKey": "value" } for a session’s temporary storage.
sessionState: Record<string, unknown>
A key-value record of the client’s session state.
Represents persistent state data (e.g., configuration, current step) for the client, akin to IState’s role in tracking agent state in ClientAgent.
Example: { "step": 3, "active": true } for a session’s current status.
executionCount: number
The number of executions performed by this client within the process. Counts operations like command runs (e.g., ClientAgent.execute) or tool calls, contributing to the process’s totalExecutionCount. Example: 10 for a client that executed 10 commands.
executionInputTotal: number
The total input size processed during executions, in a numeric unit (e.g., bytes, characters). Measures the cumulative input data (e.g., incoming messages in ClientAgent.execute), useful for assessing data throughput. Example: 1024 for 1KB of total input across executions.
executionOutputTotal: number
The total output size generated during executions, in a numeric unit (e.g., bytes, characters). Measures the cumulative output data (e.g., results in ClientAgent._emitOutput), indicating response volume. Example: 2048 for 2KB of total output.
executionInputAverage: number
The average input size per execution, in a numeric unit (e.g., bytes, characters). Calculated as executionInputTotal divided by executionCount, providing a normalized input metric. Example: 102.4 for an average of 102.4 bytes per execution.
executionOutputAverage: number
The average output size per execution, in a numeric unit (e.g., bytes, characters). Calculated as executionOutputTotal divided by executionCount, offering insight into typical output size. Example: 204.8 for an average of 204.8 bytes per execution.
executionTimeTotal: string
The total execution time for the client, formatted as a string. Represents the cumulative duration of all executions (e.g., from incoming to output in ClientAgent.execute), typically in a readable format like "300ms" or "1.5s". Contributes to the process’s totalResponseTime.
executionTimeAverage: string
The average execution time per execution, formatted as a string. Calculated as executionTimeTotal divided by executionCount, providing a normalized latency metric (e.g., "30ms" per execution). Helps evaluate client-specific performance efficiency.