Downgrading to version 10.10.5

Help
1

Hi is it possible to migrate:down the database to a specific version? Specifically I want to downgrade 1 of my apps (new version, just upgraded directus) from 11.1.0 to 10.10.5 because 11.1.0 has horrible performance on my shared hosting and 10.10.5 works well

Also does anyone know how to find directus docs for old versions (particularly directus 10.10.5, @ directus/api 19.0.2, @ directus/sdk 15.1.0) especially JS SDK docs?

2

Hello @customautosys

First please create a backup and on that, try using the old 10.10.5 docker image and do the docker compose up --build -d, see if it works.

Go to this link https://docs.directus.io/docs/api.html then click on the Take me home button and you will have the old document

Note: Opening the link directly will redirect to the new docs https://docs.directus.io/

3

I would rather try to figure out what’s causing the issues on the VPS. It might not be related to Directus at all.
Did you check Directus logs which queries cause the performance issues?

Edit: Is it even about Directus or just issues with the Directus SDK?

4

Hi thanks both for your advice.

I managed to downgrade to 10.10.5. The speed is somewhat faster but still too slow.

Unfortunately I’m not using a VPS but shared hosting. My main website at https://customautosys.com is significantly faster than this web app on https://customautosys.com/caselist/web and both are now running the same directus version 10.10.5 with the same @ directus/sdk and @ directus/api versions, so I’m not sure what is wrong. Both are also using Postgresql.

5

I see, so we can’t rule out any possible routing issues from this side.
Can you access /server/health endpoint of your directus instance?

6

I got this from /server/health:

{
	"status": "warn",
	"releaseId": "10.10.7",
	"serviceId": "customautosys",
	"checks": {
		"postgres:responseTime": [
			{
				"status": "warn",
				"componentType": "datastore",
				"observedUnit": "ms",
				"observedValue": 661.302,
				"threshold": 150
			}
		],
		"postgres:connectionsAvailable": [
			{
				"status": "ok",
				"componentType": "datastore",
				"observedValue": 5
			}
		],
		"postgres:connectionsUsed": [
			{
				"status": "ok",
				"componentType": "datastore",
				"observedValue": 3
			}
		],
		"storage:local:responseTime": [
			{
				"status": "warn",
				"componentType": "objectstore",
				"observedValue": 2432.579,
				"observedUnit": "ms",
				"threshold": 750
			}
		],
		"email:connection": [
			{
				"status": "ok",
				"componentType": "email"
			}
		]
	}
}

It seems like the problem is the data store and the object store, i.e. postgres. However, I’m using the same postgresql backend as my main website, just with a different database.

Also what’s the difference between objectstore and datastore?

Any idea how I might be able to fix this?

I’ve also taken a look at phpPgAdmin and both databases are 10 MB so I don’t really see what is significantly different between the databases.

Also I’ve checked the pg npm package version in both my main website and this web-app and both are 8.11.4.

This is what I get for my main website:

{
	"status": "ok",
	"releaseId": "10.10.7",
	"serviceId": "customautosys",
	"checks": {
		"postgres:responseTime": [
			{
				"status": "ok",
				"componentType": "datastore",
				"observedUnit": "ms",
				"observedValue": 3.572,
				"threshold": 150
			}
		],
		"postgres:connectionsAvailable": [
			{
				"status": "ok",
				"componentType": "datastore",
				"observedValue": 5
			}
		],
		"postgres:connectionsUsed": [
			{
				"status": "ok",
				"componentType": "datastore",
				"observedValue": 0
			}
		],
		"storage:local:responseTime": [
			{
				"status": "ok",
				"componentType": "objectstore",
				"observedValue": 8.56,
				"observedUnit": "ms",
				"threshold": 750
			}
		],
		"email:connection": [
			{
				"status": "ok",
				"componentType": "email"
			}
		]
	}
}
7

The datastore refers to your database system (e.g., Postgres), while the object store is used for binary file storage (e.g., local disk storage).

Might be an I/O or routing issue on your provider side.

I would recommend filing a ticket with your provider and waiting for their response.

8

Hi Nik, I’m so sorry for all the trouble.

Actually the problem was that I forgot to turn off the JavaScript obfuscator for my server side code, which resulted in the code lagging the server really badly.

I’ve modified that and also did a reinstall of my node_modules, and now it’s working.

Thanks for your help anyway!