RDS Serveless V1 Timeouts

0

Since a few hours ago, we have been getting seemingly random timeouts from our RDS Serverless 1 cluster. They are very simple single table queries, eg looking up a user. A query can fail, then succeed as normal when running immediately again. Successful queries don't appear to be any slower than usual.

We have been using serverless since its launch, and have not experienced this before. Queries are called from different EC2 instances, they all exhibit the same problem, so I don't think the problem is there. The servers have been running for months, no recent changes.

The timeouts are similar to what happens when I try to access the database from a different VPC without granting access. Maybe the connections are being routed through a different something??

Gary

  • Setting up some tests and I am totally lost. Simple php page on a new server setup:

    $pdo = new PDO("mysql:host=$servername", $username, $password);

    Keep reloading page, connects OK about 4/5 times.

已提问 2 年前205 查看次数
2 回答
0

Any chance your usage of the database is less than it used to be, and you're scaling down to 0? Cold starts are pretty long - usually over 30 seconds so do cause timeouts.

专家
已回答 2 年前
  • No, it is set to a minimum level 2 capacity units.

  • For example, we have a button on our dashboard to reload a list of the user documents. Click that 4 times in a row and it will work, then the 5th one will fail. And may then work on the next 2, then timeout again. All running the same code on the same server.

0

And as quickly as the timeouts started, they stopped about 10 hours after they started. I did start a new database and the servers had no trouble connecting to it, so I was about to do a snapshot / restore to a new serverless database, when the old database started working again. No idea why.

已回答 2 年前

您未登录。 登录 发布回答。

一个好的回答可以清楚地解答问题和提供建设性反馈,并能促进提问者的职业发展。

回答问题的准则