feat: implement model lock functionality for connection management

This commit is contained in:
decolua
2026-02-27 10:29:11 +07:00
parent 1f4423d444
commit 25c2ad7360
6 changed files with 219 additions and 205 deletions

View File

@@ -1,67 +1,11 @@
import { getProviderConnections, validateApiKey, updateProviderConnection, getSettings } from "@/lib/localDb";
import { isAccountUnavailable, getUnavailableUntil, getEarliestRateLimitedUntil, formatRetryAfter, checkFallbackError } from "open-sse/services/accountFallback.js";
import { formatRetryAfter, checkFallbackError, isModelLockActive, buildModelLockUpdate, buildClearModelLocksUpdate, getEarliestModelLockUntil } from "open-sse/services/accountFallback.js";
import { resolveProviderId } from "@/shared/constants/providers.js";
import * as log from "../utils/logger.js";
// Mutex to prevent race conditions during account selection
let selectionMutex = Promise.resolve();
// ---------------------------------------------------------------------------
// Model-level rate-limit locking (in-memory)
//
// Providers like Antigravity maintain *separate* quota buckets per model
// family (e.g. Claude vs Gemini). When a 429 arrives for one model, the
// standard account-level DB lock would block ALL models on that account,
// wasting quota that is still available for other model families.
//
// This module tracks model-specific locks in memory so that only the
// affected model is skipped during account selection while the rest of the
// account's quota remains accessible.
//
// Locks are intentionally in-memory: they clear on restart and require no
// database schema migration.
// ---------------------------------------------------------------------------
/** Providers known to have independent per-model quota buckets */
const MULTI_BUCKET_PROVIDERS = new Set(["antigravity"]);
/** Map<"connectionId:model", expiryTimestamp> */
const modelLocks = new Map();
/** Default lock duration for model-level rate limits (5 minutes) */
const DEFAULT_MODEL_LOCK_MS = 5 * 60 * 1000;
/**
* Check whether a specific model is temporarily locked on a connection.
* Expired locks are cleaned up lazily.
*/
function isModelLocked(connectionId, model) {
if (!connectionId || !model) return false;
const key = `${connectionId}:${model}`;
const expiry = modelLocks.get(key);
if (!expiry) return false;
if (expiry > Date.now()) return true;
modelLocks.delete(key); // clean up expired
return false;
}
/**
* Lock a model on a specific connection for `durationMs` milliseconds.
*/
function lockModel(connectionId, model, durationMs) {
if (!connectionId || !model) return;
const key = `${connectionId}:${model}`;
modelLocks.set(key, Date.now() + durationMs);
log.warn("AUTH", `Model lock: ${model} on ${connectionId.slice(0, 8)} for ${Math.round(durationMs / 1000)}s`);
}
/**
* Check whether a provider uses per-model quota buckets.
*/
function isMultiBucketProvider(provider) {
return MULTI_BUCKET_PROVIDERS.has(provider);
}
/**
* Get provider credentials from localDb
* Filters out unavailable accounts and returns the selected account based on strategy
@@ -85,50 +29,35 @@ export async function getProviderCredentials(provider, excludeConnectionId = nul
log.debug("AUTH", `${provider} | total connections: ${connections.length}, excludeId: ${excludeConnectionId || "none"}, model: ${model || "any"}`);
if (connections.length === 0) {
// Check all connections (including inactive) to see if rate limited
const allConnections = await getProviderConnections({ provider: providerId });
log.debug("AUTH", `${provider} | all connections (incl inactive): ${allConnections.length}`);
if (allConnections.length > 0) {
const earliest = getEarliestRateLimitedUntil(allConnections);
if (earliest) {
log.warn("AUTH", `${provider} | all ${allConnections.length} accounts rate limited (${formatRetryAfter(earliest)})`);
return { allRateLimited: true, retryAfter: earliest, retryAfterHuman: formatRetryAfter(earliest) };
}
log.warn("AUTH", `${provider} | ${allConnections.length} accounts found but none active`);
allConnections.forEach(c => {
log.debug("AUTH", ` → ${c.id?.slice(0, 8)} | isActive=${c.isActive} | rateLimitedUntil=${c.rateLimitedUntil || "none"} | testStatus=${c.testStatus}`);
});
}
log.warn("AUTH", `No credentials for ${provider}`);
return null;
}
// Filter out unavailable accounts and excluded connection
const multiBucket = isMultiBucketProvider(provider);
// Filter out model-locked and excluded connections
const availableConnections = connections.filter(c => {
if (excludeConnectionId && c.id === excludeConnectionId) return false;
if (isAccountUnavailable(c.rateLimitedUntil)) return false;
if (multiBucket && model && isModelLocked(c.id, model)) return false;
if (isModelLockActive(c, model)) return false;
return true;
});
log.debug("AUTH", `${provider} | available: ${availableConnections.length}/${connections.length}`);
connections.forEach(c => {
const excluded = excludeConnectionId && c.id === excludeConnectionId;
const rateLimited = isAccountUnavailable(c.rateLimitedUntil);
const modelLocked = multiBucket && model && isModelLocked(c.id, model);
if (excluded || rateLimited || modelLocked) {
log.debug("AUTH", ` → ${c.id?.slice(0, 8)} | ${excluded ? "excluded" : ""} ${rateLimited ? `rateLimited until ${c.rateLimitedUntil}` : ""} ${modelLocked ? `modelLocked(${model})` : ""}`);
const locked = isModelLockActive(c, model);
if (excluded || locked) {
const lockUntil = getEarliestModelLockUntil(c);
log.debug("AUTH", ` → ${c.id?.slice(0, 8)} | ${excluded ? "excluded" : ""} ${locked ? `modelLocked(${model}) until ${lockUntil}` : ""}`);
}
});
if (availableConnections.length === 0) {
const earliest = getEarliestRateLimitedUntil(connections);
// Find earliest lock expiry across all connections for retry timing
const lockedConns = connections.filter(c => isModelLockActive(c, model));
const expiries = lockedConns.map(c => getEarliestModelLockUntil(c)).filter(Boolean);
const earliest = expiries.sort()[0] || null;
if (earliest) {
// Find the connection with the earliest rateLimitedUntil to get its error info
const rateLimitedConns = connections.filter(c => c.rateLimitedUntil && new Date(c.rateLimitedUntil).getTime() > Date.now());
const earliestConn = rateLimitedConns.sort((a, b) => new Date(a.rateLimitedUntil) - new Date(b.rateLimitedUntil))[0];
log.warn("AUTH", `${provider} | all ${connections.length} active accounts rate limited (${formatRetryAfter(earliest)}) | lastErrorCode=${earliestConn?.errorCode}, lastError=${earliestConn?.lastError?.slice(0, 50)}`);
const earliestConn = lockedConns[0];
log.warn("AUTH", `${provider} | all ${connections.length} accounts locked for ${model || "all"} (${formatRetryAfter(earliest)}) | lastError=${earliestConn?.lastError?.slice(0, 50)}`);
return {
allRateLimited: true,
retryAfter: earliest,
@@ -137,15 +66,6 @@ export async function getProviderCredentials(provider, excludeConnectionId = nul
lastErrorCode: earliestConn?.errorCode || null
};
}
if (multiBucket && model) {
log.warn("AUTH", `${provider} | all accounts model-locked for ${model}`);
return {
allRateLimited: true,
retryAfter: new Date(Date.now() + 60000).toISOString(),
retryAfterHuman: "reset after 1m",
lastError: `All accounts rate limited for model ${model}`
};
}
log.warn("AUTH", `${provider} | all ${connections.length} accounts unavailable`);
return null;
}
@@ -209,7 +129,8 @@ export async function getProviderCredentials(provider, excludeConnectionId = nul
// Include current status for optimization check
testStatus: connection.testStatus,
lastError: connection.lastError,
rateLimitedUntil: connection.rateLimitedUntil
// Pass full connection for clearAccountError to read modelLock_* keys
_connection: connection
};
} finally {
if (resolveMutex) resolveMutex();
@@ -217,9 +138,8 @@ export async function getProviderCredentials(provider, excludeConnectionId = nul
}
/**
* Mark account as unavailable — reads backoffLevel from DB, calculates cooldown with exponential backoff, saves new level.
* For multi-bucket providers (e.g. Antigravity), 429 errors lock only the specific model in memory
* rather than the entire account in the database.
* Mark account+model as unavailable — locks modelLock_${model} in DB.
* All errors (429, 401, 5xx, etc.) lock per model, not per account.
* @param {string} connectionId
* @param {number} status - HTTP status code from upstream
* @param {string} errorText
@@ -228,7 +148,6 @@ export async function getProviderCredentials(provider, excludeConnectionId = nul
* @returns {{ shouldFallback: boolean, cooldownMs: number }}
*/
export async function markAccountUnavailable(connectionId, status, errorText, provider = null, model = null) {
// Read current connection to get backoffLevel
const connections = await getProviderConnections({ provider });
const conn = connections.find(c => c.id === connectionId);
const backoffLevel = conn?.backoffLevel || 0;
@@ -236,17 +155,11 @@ export async function markAccountUnavailable(connectionId, status, errorText, pr
const { shouldFallback, cooldownMs, newBackoffLevel } = checkFallbackError(status, errorText, backoffLevel);
if (!shouldFallback) return { shouldFallback: false, cooldownMs: 0 };
if (isMultiBucketProvider(provider) && status === 429 && model) {
const lockDuration = cooldownMs > 0 ? cooldownMs : DEFAULT_MODEL_LOCK_MS;
lockModel(connectionId, model, lockDuration);
return { shouldFallback: true, cooldownMs: 0 };
}
const rateLimitedUntil = getUnavailableUntil(cooldownMs);
const reason = typeof errorText === "string" ? errorText.slice(0, 100) : "Provider error";
const lockUpdate = buildModelLockUpdate(model, cooldownMs);
await updateProviderConnection(connectionId, {
rateLimitedUntil,
...lockUpdate,
testStatus: "unavailable",
lastError: reason,
errorCode: status,
@@ -254,6 +167,9 @@ export async function markAccountUnavailable(connectionId, status, errorText, pr
backoffLevel: newBackoffLevel ?? backoffLevel
});
const lockKey = Object.keys(lockUpdate)[0];
log.warn("AUTH", `${connectionId.slice(0, 8)} locked ${lockKey} for ${Math.round(cooldownMs / 1000)}s [${status}]`);
if (provider && status && reason) {
console.error(`❌ ${provider} [${status}]: ${reason}`);
}
@@ -263,21 +179,26 @@ export async function markAccountUnavailable(connectionId, status, errorText, pr
/**
* Clear account error status (only if currently has error)
* Optimized to avoid unnecessary DB updates
* Clears all modelLock_* fields and resets error state.
*/
export async function clearAccountError(connectionId, currentConnection) {
// Only update if currently has error status
const hasError = currentConnection.testStatus === "unavailable" ||
currentConnection.lastError ||
currentConnection.rateLimitedUntil;
// Support both direct connection object and credentials wrapper
const conn = currentConnection._connection || currentConnection;
const now = Date.now();
// Collect all modelLock_* keys (both active and expired)
const allLockKeys = Object.keys(conn).filter(k => k.startsWith("modelLock_"));
const hasError = conn.testStatus === "unavailable" || conn.lastError || allLockKeys.length > 0;
if (!hasError) return; // Skip if already clean
// Clear all modelLock_* keys (lazy cleanup of expired ones included)
const clearLocks = Object.fromEntries(allLockKeys.map(k => [k, null]));
await updateProviderConnection(connectionId, {
...clearLocks,
testStatus: "active",
lastError: null,
lastErrorAt: null,
rateLimitedUntil: null,
backoffLevel: 0
});
log.info("AUTH", `Account ${connectionId.slice(0, 8)} error cleared`);