Performance Issues with Self-Hosted Zitadel on Heroku (Slow Login/Callback Redirects)
Original message was deleted
FOh well 20s is really a little of 
Can you share infos about your setup, like CPU/mem sizes and your zitadel config?
Do you see in heroku some insights from the DB?
Can you share infos about your setup, like CPU/mem sizes and your zitadel config?
Do you see in heroku some insights from the DB?
FThat should be enough for zitadel, how does the DB look like?
Do you have any data there?
Do you have any data there?
FYeah but are there metrics you can share?
FWhat happens if you delete? Can you get me the version that is running (zitadel does print that in the startup message) ?
ZITADEL_PROJECTIONS_TRANSACTIONDURATION : 2000ms
ZITADEL_EVENTSTORE_PUSHTIMEOUT: 60s
ZITADEL_PROJECTIONS_TRANSACTIONDURATION : 2000ms
ZITADEL_EVENTSTORE_PUSHTIMEOUT: 60s
FOk, to me this looks like the DB is somewhat slow, but its hard to tell.
FMust be a weird issue between heroku and zitadel 
FWhat plan is the database on?
FOne thing that I could see is maybe the amount of connections, or something like that
FThe usual start for a production we recommend is a 4cpu/16gb postgres 15 or 16.
That can handle up to 500/rps depending on the api call
That can handle up to 500/rps depending on the api call
FHappy to help
FHeya, certainly happy to help.
Which endpoint you see slow latency ATM?
Happy to peak into our data and check if that is sane...
On a general note, Zitadel most of the times has bad latency when it needs to wait for a DB connection.
So either pooling connections or having >20 of them ready certainly helps
Which endpoint you see slow latency ATM?
Happy to peak into our data and check if that is sane...
On a general note, Zitadel most of the times has bad latency when it needs to wait for a DB connection.
So either pooling connections or having >20 of them ready certainly helps