Skip to content
By using AWS re:Post, you agree to the AWS re:Post Terms of Use
AWS re:Postre:Post
About re:Post

Aurora Serverless v2 waking up every ~4 hours without client connections

3

Hi all,

Enter image description here

I have an Aurora Serverless v2 RDS cluster (Aurora Mysql) that I mainly use for a small side-project in development. Normally it behaves as expected — scaling to 0 after ~5 minutes of SQL inactivity — but starting about two weeks ago, July 30th, it began waking up every ~4 hours for a very brief period before going idle again.

  • Logs show no unusual activity during these wake-up windows.
  • No new client connections are recorded — only rdsadmin and rdssleepagent.
  • The issue persists even on a freshly cloned instance left untouched overnight. Also, a brand new cluster I spun just for testing purposes shows same wake-up symptoms.
  • Upgrading or remaining on the same engine version makes no difference (tried v8.08.*, v8.09.*, v8.10.*)
  • Performance Insights: Off
  • Database Insights: Standard
  • Audit, general, and error logs are being shipped to CloudWatch.
  • Multi-AZ: No; Role: regional cluster

RDS events wake-up reason: "Initiated resume for the DB instance: [redacted] due to database background activity"

Has anyone seen Aurora Serverless v2 resume on a fixed interval without any external connections? Could this be tied to internal maintenance tasks or a configuration I’m overlooking?

Thanks in advance for any insights. Enter image description here

  • MattYeayea
    a year ago

    I'm experiencing the same on MySQL. I want an answer!! I tried disabling 'event_scheduler and innodb_stats_auto_recalc' in a custom param group but no luck there

  • John W
    a year ago

    I am also seeing this issue. Started July 29th across 10 different instances.

Topics
Database
Tags
DatabaseAmazon AuroraAmazon Aurora Serverless
Language
English

asked a year ago879 views

3 Answers
3
Accepted Answer

I was able to locate the issue for my Aurora Postgres instances by checking Cloudtrail. There was an external job that made a DescribeDBClusterBacktracks call that was waking up the instances.

Revision 1: Correcting this did not resolve the issue, I am still investigating.

Revision 2: We have identified the root cause and are actively working on implementing a fix.

Revision 3: The fix is currently in the deployment process and is expected to be completed by September 30th.

answered a year ago

  • MattYeayea
    a year ago

    Did you try disabling this feature? Would this work? "aws rds modify-db-cluster --db-cluster-identifier your-cluster-identifier --backtrack-window 0"

  • John W
    a year ago

    I set a policy for the role that was making the call to deny. After making this change though it is still waking up (it appears to also correlate to automated backup events). I have raised a support ticket with AWS.

  • MattYeayea
    a year ago

    John, would you follow up here after your ticket is resolved? I'll reach out as well if I don't hear back. My company is trying to spin up Aurora clusters around the world and it would be nice if they could idle at minimal cost

  • John W
    a year ago

    I wouldn't consider the ticket 'resolved' yet, but they have followed up indicating that they have identified the root cause and are actively working on implementing a fix.

  • kaimoeller
    a year ago

    Any update on this? I am still getting >10mins unwanted up time every 4 hours on my instance(s) causing cost for no value at all.

2

I am experiencing the same issue on my Aurora Postgres v2 instances. It started July 28th without anything I have changed or even touched the instances. That causes a significant increase in cost. Help is appreciated.

answered a year ago

-3

The behavior you're experiencing with your Aurora Serverless v2 MySQL database resuming every ~4 hours despite no apparent user activity is not uncommon. The logs indicate that the resume is happening due to "database background activity," which can be triggered by several factors:

  1. Internal processes and maintenance: Aurora may perform various background tasks such as upgrades, configuration changes, or log file downloads that can trigger resuming.

  2. Continuous backup processes: If the Aurora storage system is still applying storage changes to catch up to the current point in time, this can cause the instance to resume.

  3. Engine-specific scheduled jobs: MySQL may have internal processes that periodically activate, causing the instance to resume.

  4. Log shipping to CloudWatch: Since you mentioned that audit, general, and error logs are being shipped to CloudWatch, this periodic activity could be triggering the resume events.

  5. Database Insights: Your "Standard" Database Insights setting might be collecting metrics at regular intervals, potentially causing the database to resume.

To troubleshoot this issue:

  • Check your database logs more thoroughly for patterns or recurring events coinciding with the resume times
  • Review your DB parameter group settings for configurations that might trigger periodic activities
  • Examine the exact timing of the 4-hour intervals to see if they align with any scheduled AWS maintenance windows
  • Consider setting up more detailed CloudWatch alarms to capture any subtle patterns in resource utilization just before the resumes

While the auto-pause feature helps reduce costs, it's designed to balance cost savings with availability. For a development environment that's truly idle for extended periods, you might consider manually stopping the instance when not in use, though this would result in longer startup times when needed again.
Sources
Aurora serverless keeps resuming every hour | AWS re:Post
Aurora Serverless v2 resumes without user activity | AWS re:Post
Managing Aurora Serverless v2 DB clusters - Amazon Aurora

answered a year ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Relevant content