Show role when choosing database

[slootjes]

When using Port 7777 on an RDS cluster with multiple instances the output looks something like this:

Listing databases.
Which database would you like to connect to?

[1] some-service-databaseinstance1-azznqqajdhkm
[2] some-service-databaseinstance2-cqtwprntcl09

but based on this names I don't know which of the 2 is a reader or a writer. When making changes I obviously need a writer and while just doing some queries I prefer using a reader instance. Currently I always need to login to AWS to see which one I need. It would be very nice to see the role of the instance in this selection, perhaps like this:

Listing databases.
Which database would you like to connect to?

[1] some-service-databaseinstance1-azznoqajd0km (Writer instance)
[2] some-service-databaseinstance2-cqtwprbtdl09 (Reader instance)

While writing this, maybe also add the instance size to it as well:

Listing databases.
Which database would you like to connect to?

[1] some-service-databaseinstance1-azznoqajd0km (Writer instance / Serverless v2 (0.5 - 16 ACUs))
[2] some-service-databaseinstance2-cqtwprbtdl09 (Reader instance / tg4.medium)

What do you think?

[mnapoli]

Sounds like a great improvement idea, thanks for opening this. Naive question: is the reader/writer status exposed via the RDS API?

In any case I'm taking note of the idea (no promise on an implementation date ^^)

[slootjes]

The information is returned in the describe db clusters API endpoint which is already used as far as I know by this app, this is the relevant output for my cluster:

<DBClusterMembers>
	<DBClusterMember>
		<DBInstanceIdentifier>some-service-databaseinstance1-azznoqajd0km</DBInstanceIdentifier>
		<DBClusterParameterGroupStatus>in-sync</DBClusterParameterGroupStatus>
		<PromotionTier>1</PromotionTier>
		<IsClusterWriter>true</IsClusterWriter>
	</DBClusterMember>
	<DBClusterMember>
		<DBInstanceIdentifier>some-service-databaseinstance2-cqtwprbtdl09</DBInstanceIdentifier>
		<DBClusterParameterGroupStatus>in-sync</DBClusterParameterGroupStatus>
		<PromotionTier>1</PromotionTier>
		<IsClusterWriter>false</IsClusterWriter>
	</DBClusterMember>
</DBClusterMembers>

The "IsClusterWriter" returns the relevant information if it's read/write or read only.

The instance type is available in the Describe Db Instances API call:

<DBInstanceClass>db.serverless</DBInstanceClass>

Maybe returning it as a table would be prettier even when adding more information.

[slootjes]

Any update on this by any chance? The same thing applies for Elasticache:

Which Elasticache instance would you like to connect to?

[1] blr12bzer1fbecf1-001
[2] blr12bzer1fbecf1-002
[3] blrhen961d1ou3l-001
[4] blrhen961d1ou3l-002

Here it doesn't even show the name, only the identifier given by AWS. The same information would be useful here, especially the description and role (primary/replica).

[slootjes]

Any news on this? Please let me know if there is a way for me to contribute, I don't think this should be difficult.

[mnapoli]

@slootjes no update at the moment.

Here is the relevant code if you want to have a go at it: https://gist.github.com/mnapoli/ea2300ed9496784ac91addefa183fd4f

The small challenge is that the prompt's selected result === what is displayed. So if we add more info to each line (e.g. the name, etc.), then when the user selects the line we get the whole string back.

[slootjes]

Thanks for this :) The easy way out would be to have something like "{option} - {extra usefull stuff}" and then split/trim on the - and use the first part?

[mnapoli]

Oh yeah why not actually! Splitting on some character could work

[slootjes]

I do see that the cluster ID being such a random name is because the ID is optional in CloudFormation and I didn't set one, this is not a bug in Port 7777. It would be very nice though to also get the description (which I did fill in) but I understand if this is a problem of my own :) The reader/writer (or primary/replace) would be most useful to know still.

[slootjes]

Any way I can help with this? While I love this tool, this is still a major annoyance every time I need to connect with an instance.

[mnapoli]

Hey @slootjes! I've published v1.1.20 (it should be available in a few minutes once the build finishes). I tried to add more details, the writer/reader was a bit trickier so I went with other options. This isn't perfect, but I hope it will help identify a bit better the right cluster.

Please let me know if that helps!

[slootjes]

Hey @mnapoli I see the update and it now renders this for me:

Which database would you like to connect to?

[1] <stack_name>-databaseinstance1-<id> (instance - aurora-postgresql - db.serverless)
[2] <stack_name>-databaseinstance2-<id> (instance - aurora-postgresql - db.serverless)

and while this is progress (thanks!), it doesn't solve my issue since I still don't know which database I need to select...

I'm really happy to help out here, let me know.

[mnapoli]

😢 ok at least we can iterate now! What fields should I render there to make it more useful?

FYI I have a DBCluster object (result of DescribeDBClusters), which field should I use instead?

[slootjes]

That API call returns me:

"DBClusterMembers": [
	{
		"DBInstanceIdentifier": "<stack>-databaseinstance1-<id>",
		"IsClusterWriter": false,
		"DBClusterParameterGroupStatus": "in-sync",
		"PromotionTier": 1
	},
	{
		"DBInstanceIdentifier": "<stack>-databaseinstance2-<id>",
		"IsClusterWriter": true,
		"DBClusterParameterGroupStatus": "in-sync",
		"PromotionTier": 1
	}
],

so the "IsClusterWriter"": false would result in "Read" and "IsClusterWriter": true in "Read/Write". The naming is up for debate of course.

[mnapoli]

If there are multiple DBClusterMembers which do you want to be used for the label? The first one in the list?

[slootjes]

I think the label is good as-is, for RDS at least 👍 For ElastiCache it's still showing me:

[1] blr12bzer1fbecf1-001
[2] blr12bzer1fbecf1-002
[3] blrhen961d1ou3l-001
[4] blrhen961d1ou3l-002

Could you list me the API calls you do for this too? Then I can have a look at the output theretoo.