fix: tune sub2api codex pool concurrency
This commit is contained in:
@@ -9,6 +9,7 @@ const configPath = rootPath("config", "platform-infra", "sub2api-codex-pool.yaml
|
||||
const parsed = Bun.YAML.parse(readFileSync(configPath, "utf8")) as {
|
||||
pool?: {
|
||||
defaultAccountCapacity?: number;
|
||||
minOwnerConcurrency?: number;
|
||||
defaultTempUnschedulable?: {
|
||||
enabled?: boolean;
|
||||
rules?: Array<{ statusCode?: number; keywords?: string[]; durationMinutes?: number }>;
|
||||
@@ -18,25 +19,35 @@ const parsed = Bun.YAML.parse(readFileSync(configPath, "utf8")) as {
|
||||
};
|
||||
|
||||
const entries = parsed.profiles?.entries ?? [];
|
||||
const hy = entries.find((entry) => entry.profile === "HY");
|
||||
const wsEnabled = entries.filter((entry) => entry.openaiResponsesWebSocketsV2Mode && entry.openaiResponsesWebSocketsV2Mode !== "off");
|
||||
const rules = parsed.pool?.defaultTempUnschedulable?.rules ?? [];
|
||||
const overload429 = rules.find((rule) => rule.statusCode === 429);
|
||||
const defaultCapacity = parsed.pool?.defaultAccountCapacity ?? 0;
|
||||
const desiredCapacity = entries.reduce((total, entry) => total + (entry.capacity ?? defaultCapacity), 0);
|
||||
const allowedWebSocketModes = new Set(["off", "ctx_pool", "passthrough"]);
|
||||
|
||||
assertCondition(parsed.pool?.defaultAccountCapacity === 5, "Codex pool default capacity should remain explicit YAML", parsed.pool);
|
||||
assertCondition(hy !== undefined, "HY profile should remain declared as a YAML-managed pool candidate", entries);
|
||||
assertCondition(hy?.capacity === 20, "HY capacity must remain explicitly YAML-controlled at the current configured value", hy);
|
||||
assertCondition(wsEnabled.length >= 2, "Responses WSv2 should be a capability set with multiple candidates", wsEnabled);
|
||||
assertCondition(parsed.pool?.defaultTempUnschedulable?.enabled === true, "temporary unschedulable policy must be enabled from YAML", parsed.pool?.defaultTempUnschedulable);
|
||||
assertCondition(overload429 !== undefined, "temporary unschedulable policy must include 429 handling", rules);
|
||||
assertCondition(overload429?.keywords?.includes("too many requests"), "429 handling must catch provider concurrency/rate-limit text", overload429);
|
||||
assertCondition((overload429?.durationMinutes ?? 0) > 0, "429 handling must cool down for a positive duration", overload429);
|
||||
assertCondition(entries.length > 0, "Codex pool must declare YAML-managed profile entries", parsed.profiles);
|
||||
assertCondition(Number.isInteger(defaultCapacity) && defaultCapacity > 0, "defaultAccountCapacity must be a positive integer", parsed.pool);
|
||||
assertCondition(entries.every((entry) => typeof entry.profile === "string" && entry.profile.length > 0), "profile entries must declare profile names", entries);
|
||||
assertCondition(entries.every((entry) => typeof entry.accountName === "string" && entry.accountName.length > 0), "profile entries must declare account names", entries);
|
||||
assertCondition(entries.every((entry) => entry.capacity === undefined || (Number.isInteger(entry.capacity) && entry.capacity > 0)), "profile capacity overrides must be positive integers when declared", entries);
|
||||
assertCondition(
|
||||
entries.every((entry) => entry.openaiResponsesWebSocketsV2Mode === undefined || entry.openaiResponsesWebSocketsV2Mode === null || allowedWebSocketModes.has(entry.openaiResponsesWebSocketsV2Mode)),
|
||||
"profile WebSocket mode overrides must use supported values when declared",
|
||||
entries,
|
||||
);
|
||||
assertCondition((parsed.pool?.minOwnerConcurrency ?? 0) >= desiredCapacity, "pool owner concurrency must not bottleneck the declared account capacity set", { minOwnerConcurrency: parsed.pool?.minOwnerConcurrency, desiredCapacity });
|
||||
if (parsed.pool?.defaultTempUnschedulable?.enabled === true) {
|
||||
assertCondition(rules.length > 0, "enabled temporary unschedulable policy must declare rules", parsed.pool?.defaultTempUnschedulable);
|
||||
assertCondition(rules.every((rule) => Number.isInteger(rule.statusCode) && (rule.statusCode ?? 0) >= 100 && (rule.statusCode ?? 0) <= 599), "temporary unschedulable rules must declare valid HTTP status codes", rules);
|
||||
assertCondition(rules.every((rule) => Array.isArray(rule.keywords) && rule.keywords.length > 0), "temporary unschedulable rules must declare non-empty keywords", rules);
|
||||
assertCondition(rules.every((rule) => Number.isInteger(rule.durationMinutes) && (rule.durationMinutes ?? 0) > 0), "temporary unschedulable rules must declare positive cooldown durations", rules);
|
||||
}
|
||||
|
||||
console.log(JSON.stringify({
|
||||
ok: true,
|
||||
checks: [
|
||||
"HY capacity is YAML-controlled",
|
||||
"WSv2 routing is represented as a multi-account capability set",
|
||||
"temporary unschedulable rules are YAML-controlled",
|
||||
"routing config is schema-valid without profile-specific test gates",
|
||||
"pool owner concurrency covers the YAML account capacity set",
|
||||
"optional WebSocket mode overrides use supported values",
|
||||
"temporary unschedulable rules are structurally valid when enabled",
|
||||
],
|
||||
}));
|
||||
|
||||
Reference in New Issue
Block a user