rds for mysql with high disk io await ?

Question

**situation:**
I have two rds for mysql databases with same configuration, and they have similar workloads,  I split my traffic evenly into these two databases. But the performance varies a lot in db load. After some digging, I found that the disk io await varies a lot between these two dbs.   My guess is that maybe there is something wrong with the database disk.  does someone have the same situation?
![Enter image description here](/media/postImages/original/IM0vbZ_m83RNSG81U0RZSkSw)

![Enter image description here](/media/postImages/original/IMI4Zx3wXhT0uPkJqtpdIO3A)

Answer

I think the issue here is that you are running out of available IOPS capacity on the ec2 instance class you've selected. Eg, this instance has a max iops of 20k.

```
$ aws ec2 describe-instance-types --instance-types m6g.4xlarge --query "InstanceTypes[*].[InstanceType, EbsInfo.EbsOptimizedSupport, EbsInfo.EbsOptimizedInfo]"
[
    [
        "m6g.4xlarge",
        {
            "BaselineBandwidthInMbps": 4750,
            "BaselineThroughputInMBps": 593.75,
            "BaselineIops": 20000,
            "MaximumBandwidthInMbps": 4750,
            "MaximumThroughputInMBps": 593.75,
            "MaximumIops": 20000
        }
    ]
]

Answer

After I changed the disk configuration,  iops from 12000 to 13000, I did this modification to triggle the storage optimization action. And the disk io await goes to normal state.  Very strange.

rds for mysql with high disk io await ?

相關內容