Open hagaibarel opened 10 months ago
Oh damn, did it work on 0.32.5? I think we unified cache configs but frontend parses stricter then bucket caches.
I haven't tried in with 0.32.5, AFAIK this was added in https://github.com/thanos-io/thanos/pull/6773 which was part of 0.33.0
Hello @hagaibarel ,
The query-frontend does not have a ttl config for its cache, but an expiration :/ https://thanos.io/tip/components/query-frontend.md/#redis
Thanks @ahurtaud for clearing that up, are you aware of any plans to unify the configuration?
Thanks @ahurtaud for clearing that up, are you aware of any plans to unify the configuration?
no, not aware. it should be done I think to avoid confusion
I agree! Do you mind opening a pull request for it?
I dont feel I have the time currently to do it sorry. Also, that would mean a breaking change now that 0.33 is released? I am not sure we want to do it right now. Maybe a refactor of the cache configs so they use the same code (query-frontend and store).
I dont feel I have the time currently to do it sorry. Also, that would mean a breaking change now that 0.33 is released? I am not sure we want to do it right now. Maybe a refactor of the cache configs so they use the same code (query-frontend and store).
I would accept both and then error if both were provided i think. Ill see if i can muster some time on the weekend!
We should definitely unify configurations at some point. The current QFE cache config is from Cortex code so not the same cache client code.
Thanos, Prometheus and Golang version used:
Thanos -
quay.io/thanos/thanos:v0.33.0
(deployed on kubernetes) Prometheus / Golang = N/AObject Storage Provider: GCS
What happened:
Specifying
ttl
in theredis
cache config causes query fronted to fail to start with an exceptionWhat you expected to happen:
Normal startup
How to reproduce it (as minimally and precisely as possible):
Start query frontend with the following redis config:
Full logs to relevant components: