I'm encountering an issue with the rate limit plugin in my project. Specifically, when using the resolveLimits function with the sequence Limit::allow(1)->everySeconds(1)->sleep(), the plugin doesn't consistently adhere to the specified rate limit of one request per second. Occasionally, it exceeds this limit and makes multiple requests within a single second.
Upon investigating the plugin's code, I noticed a potential issue within the handleExceededLimit function. There seems to be a scenario where the $limit->getRemainingSeconds value reaches zero, possibly leading to the delay not being properly applied. While I can't confirm if this is the root cause, I made a modification to this method within my connector, and it appears to have mitigated the issue.
In production, with the limit of one request per second and without changing anything in the handleExceededLimit method:
In localhost, setting the delay in handleExceededLimit as 0:
In localhost, setting the delay in handleExceededLimit to 1 if it $limit->getRemainingSeconds reaches 0:
My current connector:
<?php
namespace Connector;
use App\Exceptions\UnauthorizedSetup;
use App\Http\Middleware\Saloon\ResponseLogger;
use App\Models\Store;
use App\Traits\InteractsWithSaloonResponses;
use GuzzleHttp\RequestOptions;
use GuzzleHttp\TransferStats;
use Illuminate\Support\Facades\Cache;
use MyIntegration\Http\Pagination\Paged;
use Saloon\Http\Connector;
use Saloon\Http\PendingRequest;
use Saloon\Http\Request;
use Saloon\Http\Response;
use Saloon\PaginationPlugin\Contracts\HasPagination;
use Saloon\RateLimitPlugin\Contracts\RateLimitStore;
use Saloon\RateLimitPlugin\Limit;
use Saloon\RateLimitPlugin\Stores\LaravelCacheStore;
use Saloon\RateLimitPlugin\Stores\MemoryStore;
use Saloon\RateLimitPlugin\Traits\HasRateLimits;
use Saloon\Traits\Plugins\AcceptsJson;
use Saloon\Traits\Plugins\AlwaysThrowOnErrors;
use Saloon\Traits\Plugins\HasTimeout;
use Symfony\Component\HttpFoundation\Response as ResponseCode;
class IntegrationConnector extends Connector implements HasPagination
{
use AcceptsJson;
use AlwaysThrowOnErrors;
use HasRateLimits {
hasRateLimits::handleTooManyAttempts as protected traitHandleTooManyAttempts;
}
use HasTimeout;
use InteractsWithSaloonResponses;
public ?string $requestTime = null;
protected int $requestTimeout = 10;
public ?int $tries = 3;
public ?bool $useExponentialBackoff = true;
private const LIMITER = 'integration-connector';
public function __construct(private readonly Store $store)
{
$oAuth = $this->store->oAuth()->active()->first();
if (is_null($oAuth)) {
throw new UnauthorizedSetup('Could not find active oAuth to use.', fatal: true);
}
$this->middleware()->onRequest(
static fn (PendingRequest $pendingRequest) => $pendingRequest->headers()
->add('Authorization', 'Bearer '.$oAuth->refresh()->access_token),
);
$this->middleware()->onResponse(new ResponseLogger($store));
$this->retryInterval = app()->environment('production') ? 2000 : null;
}
public function resolveBaseUrl(): string
{
return 'integrationUrl';
}
protected function defaultConfig(): array
{
$time = &$this->requestTime;
return [
RequestOptions::VERIFY => false,
RequestOptions::ON_STATS => function (TransferStats $stats) use (&$time) {
$time = (string) $stats->getTransferTime();
},
];
}
public function paginate(Request $request): Paged
{
return new Paged($this, $request);
}
protected function resolveLimits(): array
{
return [
Limit::allow(1)->everySeconds(1)->name('second')->sleep(),
Limit::allow(60)->everyMinute()->name('minute')->sleep(),
Limit::allow(290)->everyFiveMinutes()->name('five-minutes')->sleep(),
Limit::allow(118_000)->everyDay()->name('daily'),
];
}
protected function resolveRateLimitStore(): RateLimitStore
{
if (! app()->environment('production')) {
return new MemoryStore();
}
return new LaravelCacheStore(Cache::store('redis'));
}
protected function handleTooManyAttempts(Response $response, Limit $limit): void
{
if ($response->status() === ResponseCode::HTTP_TOO_MANY_REQUESTS) {
$this->responseLogger($this->store, $response);
}
$this->traitHandleTooManyAttempts($response, $limit);
}
protected function handleExceededLimit(Limit $limit, PendingRequest $pendingRequest): void
{
if (! $limit->getShouldSleep()) {
$this->throwLimitException($limit);
}
$existingDelay = $pendingRequest->delay()->get() ?? 0;
$remainingMilliseconds = ($limit->getRemainingSeconds() === 0 ? 1 : $limit->getRemainingSeconds()) * 1000;
$pendingRequest->delay()->set($existingDelay + $remainingMilliseconds);
}
protected function getLimiterPrefix(): ?string
{
return self::LIMITER;
}
}
I'm seeking guidance on identifying the underlying cause of this inconsistency and exploring potential solutions. Could you please provide insights into what might be causing this behavior and suggest possible remedies?
I'm encountering an issue with the rate limit plugin in my project. Specifically, when using the resolveLimits function with the sequence Limit::allow(1)->everySeconds(1)->sleep(), the plugin doesn't consistently adhere to the specified rate limit of one request per second. Occasionally, it exceeds this limit and makes multiple requests within a single second.
Upon investigating the plugin's code, I noticed a potential issue within the handleExceededLimit function. There seems to be a scenario where the $limit->getRemainingSeconds value reaches zero, possibly leading to the delay not being properly applied. While I can't confirm if this is the root cause, I made a modification to this method within my connector, and it appears to have mitigated the issue.
In production, with the limit of one request per second and without changing anything in the handleExceededLimit method:
In localhost, setting the delay in handleExceededLimit as 0:
In localhost, setting the delay in handleExceededLimit to 1 if it $limit->getRemainingSeconds reaches 0:
My current connector:
I'm seeking guidance on identifying the underlying cause of this inconsistency and exploring potential solutions. Could you please provide insights into what might be causing this behavior and suggest possible remedies?