Posgres backend unusably slow

Hello,

Retool seems to be exceptionally slow with the data transfer part of postgres database queries.

For a single row of 50 columns it seems to be taking anywhere from 4-7 seconds. With ~90%+ of the time being the 'transfer data' backend.

The dataserver is in New York and shouldn't be an issue in regards to latency.

Is there some setting or something that can be altered at Retool's end for this?

It is making the apps almost unusably slow for my users.

Thank you,

image

per this retool page: App performance best practices | Retool Docs

  • Transfer data: The time taken to transfer the query and results between the Retool backend and frontend.

So does this imply that the issue is at Retool's end internally?

Hey @bg1900! Is this something that suddenly popped up as of your writing this post, or has it been going on for a while now?

Hi Kabirdas!

It has been going on for a while.

@bg1900 Is this on a self-hosted version of Retool or Retool cloud? If self-hosted, what version are you on?

The time taken to transfer the query and results between the Retool backend and frontend.
does this imply that the issue is at Retool's end internally

This does imply that the largest amount of time for this query is in the transfer of the data from the Retool backend to the frontend. Is this happening in all your postgres queries?

Can you share a sample of what one row of data might look like? ie: what data types columns are for this table etc.

This is cloud hosted.

Due to data privacy concerns I canno tshare the data. But I can tell you it is currently ~80 columns, all varchar type. Short amounts of information in each (ranging from blank/null to ~30 characters)

Please let me know if there is further information that you need

Since this is cloud hosted, if you share your subdomain or org name, I should be able to look up some logs. Feel free to DM that information if you prefer.

Is this happening in all your postgres queries? Do you also get a response size in the query processing time breakdown? And to confirm the 4.9s total time is for 1 row of data?

Screenshot 2023-10-30 at 11.13.25 AM

Hi Joe,

I would be happy to privately share that information with you / I will send that to you now.

It is happening in all postgres queries.
Retool does not appear to give me a response size.

Yes, 4.9s for 1 row of data (sometimes it is a tad faster but still many seconds)

Hi Joe,

I am sure you're busy, but if possible could you take a look at this issue? It seems like a major ongoing problem on the backend of Retool.

Would another Retool staff member be able to comment? This issue has been going on for over a month and really limits the utility of Retool made apps

Hey @bg1900, still unable to repro and we don't see any general latency spikes with Postgres resources. Can you share a sanitized sample data (1 row is sufficient) so that we can get an idea of what is going on. Would also be helpful to have the name of the resource and the name of one of the queries that is taking an extended amount of time to return so that I can dig into the logs for your instance. Thanks.

Thank you for looking into it further.
Can we message off the forum board to look into this further. I will happily provide you all the information you need.

Thanks!

@bg1900 You are welcome to DM me the information as you did before. Thanks!

1 Like

sent to you, thanks!

ah so interestingly, for those playing along at home, it seems that the issue is Retool trying to load 2 other small javascript queries at the same time.

I am not sure why Retool reports that as a delay in transfer data but the issue resolves if those other scripts are stopped.