Generate AWS RDS inventories using the AWS CLI

Picture this: you received a notification from AWS regarding upgrading a few databases in the upcoming maintenance window. However, the DBA is on PTO, and you need to share this information with up the chain of command in a spreadsheet. What do you do?

In case you already know the AWS account and the region affected, then it would be as simple as doing the following:

aws rds describe-db-clusters \
  --region='[the region goes here]' \
  --no-paginate --output text \
  --query 'DBClusters[*].[DBClusterIdentifier,Engine,EngineVersion,Endpoint,ClusterCreateTime,DBClusterArn,Status]'

If you thought about copying that output directly from the terminal by clicking and selecting, please don’t and use pbcopy instead. You can use this alias for said command if you are a GNU Linux user. In that way, it would end looking like this:

aws rds describe-db-clusters \
  --region='[the region goes here]' \
  --no-paginate --output text \
  --query 'DBClusters[*].[DBClusterIdentifier,Engine,EngineVersion,Endpoint,ClusterCreateTime,DBClusterArn,Status]' \
  | pbcopy

Then it would be a matter of opening the spreadsheet application of your choice and using the “special paste” options to fill in the values. And you are done.

What if I don’t know the region

Ah! I have been there as well. For this, we’ll use our shell scripting knowledge and the help of another AWS cli subcommand.

for region in $(aws ec2 describe-regions --output text | cut -f4); do
  echo 'Listing RDS in region: ' $region
  AWS_PAGER=cat aws rds describe-db-clusters \
    --region="$region" \
    --no-paginate --output text \
    --query 'DBClusters[*].[DBClusterIdentifier,Engine,EngineVersion,Endpoint,ClusterCreateTime,DBClusterArn,Status]'
done

On that snippet, we iterate per each enabled region in our account and then query the clusters within them. At the moment, I haven’t figured out how to insert pbcopy in the mix without requiring temporary files, so for simplicity I’ll keep it that way.

Regarding the AWS_PAGER=cat you see there, it is to force the AWS cli to simply print the output instead of opening a new buffer/pager.

A word of caution about clusters and rogue DB instances

Ideally, all our databases should run as clusters, but in the real world, you will find at least one DB instance that does not belong to any. These “cluster-less” instances won’t appear in the queries shared above, so for those scenarios, you could use this command instead when you know the region:

aws rds describe-db-instances \
  --region='[the region goes here]' \
  --output text --no-paginate \
  --query 'DBInstances[*].[DBInstanceIdentifier,Engine,EngineVersion,Endpoint.Address,InstanceCreateTime,DBInstanceArn,DBInstanceStatus]'

And this other one to iterate over all the available regions:

for region in $(aws ec2 describe-regions --output text | cut -f4); do
  echo 'Listing RDS in region: ' $region
  AWS_PAGER=cat aws rds describe-db-instances \
    --region $region \
    --output text --no-paginate \
    --query 'DBInstances[*].[DBInstanceIdentifier,Engine,EngineVersion,Endpoint.Address,InstanceCreateTime,DBInstanceArn,DBInstanceStatus]'
done

Remember that these commands will list all the database instances, including those belonging to a cluster. So you’ll need to cross check with the clusters to spot the outliners.

Problem solved, but before you go

I hope these snippets have helped you out with the not-so-imaginary situation we described at the beginning, and before you go, I wanted to share this:

  • Prefer pbcopy over clicking and copying directly from the terminal.
  • Once you have the values, remember to use the “special paste” to fill out the spreadsheet of your choice.
  • You can get more data by changing the --query flag value. You can find the available fields within the documentation for rds describe-clusters and rds describe-instances commands.
  • Personally, I prefer to set a default region before executing the snippets you found here with the aws configure command, to avoid repeating myself too often.
  • Kudos to @albarki for sharing the aws ec2 describe-regions loop snippet over here.

Happy hacking.


Last modified on 2023-01-23