Clearing the disk cache from the Thanos Store Gateway

[Vidyasagar9390]

Hi Team,

In our Thanos implementation, we use the thanos store gateway component which is having some disk for caching purpose. Currently, disk space usage reaches beyond 75%. How can we delete the old cached items from this disk to make some free space?

Thanks,
vidyasagar

[GiedriusS]

Hey, check if your compactor is running properly. This is a sign that your compactor has a huge backlog and it is not able to do its work on time.

[Vidyasagar9390]

level=info ts=2025-07-16T07:14:05.400807973Z caller=retention.go:42 msg="applying retention policy" ...
level=info ts=2025-07-16T07:14:05.410900299Z caller=compact.go:620 msg="starting compaction"

There is no backlog.
Retention is being applied.
There is no indication that compactor is causing your Store Gateway cache disk to grow.

GitHub issue #7029,can you see this
may be this one correctly pointing out my problem
what should i do

[Vidyasagar9390]

I’ve reviewed our setup and also checked the logs from our Thanos Compactor. It appears that the compactor is functioning correctly — it’s applying retention policies, running compaction cycles, and performing garbage collection as expected, without any backlog or errors. So, I believe the compactor is not the root cause of the issue in our case.

Upon further investigation, it looks like the disk usage growth is specifically related to the Store Gateway’s local index cache. Since we’re using lazy index header download, the cache builds up over time, but there’s no automatic cleanup for unused index headers unless explicitly configured

[Vidyasagar9390]

i want a suggestion for how to clear the disk cache
please do let me know if there’s anything else I should check or optimize!