5

花了很长时间才弄清楚这一点，所以这里是寻找它的其他人的解决方案。解决方案在 PHP 中实现：

$url = 'https://host.td/api/search/universal/absolute'
       . '?query=' . urlencode('field:value')                 //query which you would also perform on UI
       . '&from=' . urlencode(Carbon::createFromTimestamp(0)) // min timestamp so we get all logs
       . '&to=' . urlencode(Carbon::createFromTimestamp(NumberUtils::MAX_32_BIT_INT)) // max timestamp so we get all logs
       . '&limit=' . $this->limit                             //how many results do we want?
       . '&fields=' . urlencode('field1,field2,field3')       //which fields do we want?
       . '&filter=' . urlencode('streams:<stream_id>')        //OPTIONAL: only search in this stream
       . '&sort=' . urlencode('field:desc')                   //sort result
       . '&decorate=false';                                   //dont know whats that
$res = (new Client())->get($url, [
    // generate a token on graylog UI;
    // we use basic auth, username=the token; password: hard coded string 'token'
    'auth'    => ['<token_value>', 'token'],  
    'headers' => ['Accept' => 'application/json']             //we want a json result
]);

$json = \GuzzleHttp\json_decode($res->getBody());

奖励：如果您想按您提供的时间戳排序，请不要将其称为时间戳，因为在这种情况下使用的是 graylogs 时间戳，而不是您的时间戳。我最终在我存储的每个字段上都使用了一个前缀。

4

0 回答 0