t8y2/dbx · error · ValueError

unknown workload kind: {workload['kind']}

Error message

unknown workload kind: {workload['kind']}

What it means

execute_workload only understands workload kinds "rpc" and "paged"; anything else raises ValueError with the unknown kind. It is a strict enum-style validation on the workload dict produced by configured_workloads.

Source

Thrown at agents/drivers/argo-go/bench/agent_compare.py:491

    for _ in range(operations):
        started = time.perf_counter()
        execute_workload(process, workload, session_id)
        samples.append(elapsed_ms(started))
    return samples


def execute_workload(
    process: AgentProcess,
    workload: dict,
    agent_session_id: str = "",
) -> object:
    params = dict(workload.get("params", {}))
    if agent_session_id:
        params["agentSessionId"] = agent_session_id
    if workload["kind"] == "rpc":
        return process.call(workload["method"], params)
    if workload["kind"] != "paged":
        raise ValueError(f"unknown workload kind: {workload['kind']}")
    first = process.call(
        "execute_query_page",
        {
            "sql": workload["sql"],
            "maxRows": workload["max_rows"],
            "pageSize": workload["page_size"],
            **({"agentSessionId": agent_session_id} if agent_session_id else {}),
        },
    )
    rows = len(first.get("rows", []))
    session_id = first.get("session_id")
    has_more = first.get("has_more", False)
    try:
        while has_more:
            page = process.call(
                "fetch_query_page",
                {
                    "sessionId": session_id,

View on GitHub (pinned to c0390bff16)

Solutions

1. Change the workload's `kind` in the config to "rpc" or "paged"
2. Fix casing/typo — the comparison is exact and case-sensitive
3. If a new kind is needed, add an explicit branch in execute_workload implementing it
4. Add validation in configured_workloads so bad kinds are rejected with a list of valid kinds

Example fix

// before
{"kind": "page", "sql": "SELECT ..."}
// after
{"kind": "paged", "sql": "SELECT ...", "max_rows": 1000, "page_size": 100}

VALID_KINDS = {"rpc", "paged"}
for w in workloads:
    if w.get("kind") not in VALID_KINDS:
        raise ValueError(f"{w.get('kind')!r} not in {sorted(VALID_KINDS)}")

def is_valid_workload(w: dict) -> bool:
    return w.get("kind") in ("rpc", "paged")

try:
    rows = execute_workload(process, workload, agent_session_id)
except ValueError as e:
    if str(e).startswith("unknown workload kind"):
        print(f"skipping workload: {e}")
    else:
        raise

When it happens

Trigger: A workload dict whose `kind` is neither "rpc" nor "paged" reaches execute_workload — via benchmark_workload, benchmark_concurrency, or concurrency_worker — typically from a hand-edited workload config or a new kind added in config without a handler in execute_workload.

Common situations: Typo like "page" or "Paged" (case-sensitive check); user added a new workload kind (e.g. "stream") in config but only edited the config, not the dispatcher; older workload JSON format migrated incorrectly.

Understand the failure class

Background: Invalid enum value errors: "Unknown type", "Invalid scope", "must be one of" — when a string is not on the library's allowed list — this error's family across 23 libraries.

Related errors

• Agent runtime thread limits must be positive
• Unsupported H2 driver profile: " + profile
• Unsupported peek startPosition: ${value}
• MongoDB explain verbosity must be queryPlanner, executionSta
• JDBC URL is required.

AI-assisted analysis of t8y2/dbx@c0390bff16 (2026-09-05). Data as JSON: /api/errors/8825434eb297eee5. Report an issue: GitHub.