Repository navigation
Payload Cloud Caching Clarification #18573
Unanswered
rohith-goose
asked this question in
Q&A
Replies: 1 comment
|
thats the classic per-instance cache problem — each payload instance caches queries in its own memory, so a write on instance A leaves instance B serving stale content. fix: use a shared cache across instances. payload supports redis as the cache handler — point every instance at the same redis and invalidations become visible everywhere. check the cache config in the docs, swap the default in-memory handler for redis. if youre already on redis and still seeing staleness, look at ttl settings — something may be caching longer than expected. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
I was updating content on my Payload Cloud-hosted website and noticed that the update was visible only to some users and not all of them. Furthermore, it was working for me in some tabs, while other tabs were serving stale content.
I asked Claude to check if the server has multiple instances, and it says that the server is hosted on DigitalOcean and appears to have multiple instances.
Payload Cloud does not provide caching at the moment, as far as I know. I think it is critical, and they should have added an external caching layer.
Can anyone let me know if my analysis is correct and if I need to implement a shared cache handler for this?
Using the Payload PRO plan
Cheers, and thank you!
All reactions