"Mark all observations as seen" seems slower in production than other environments

riparias / early-alert-webapp

The source code of the LIFE RIPARIAS early alert web application

https://alert.riparias.be

MIT License

0 stars 0 forks source link

"Mark all observations as seen" seems slower in production than other environments #11

Open niconoe opened 2 years ago

niconoe commented 2 years ago

It has resulted in a timeout (after 3 minutes) when marking all observations as seen for me after initial data load (the configuration is now updated so timeout comes after 10 minutes)
I guess it's slower because of the RDS performance (vs a local Postgis). It would be great to checkk if there's no basic DB optimization (indexes?) that could be done to improve the performances)

niconoe commented 2 years ago

I have manually checked can confirm the timeout increases allow to flag all observations as seen without problem now. It would still be great to improve performances for this (quite heavy) operation.

niconoe commented 2 years ago

Observation import is also visibly slower, but not "too slow" IMHO. I guess RDS performance with the actual plan is lower than with a manually installed Postgres instance on EC2.

niconoe commented 11 months ago

That's confirmed, it would actually be nice to update the RIPARIAS prod instance to a larger database, as already discussed/requested.