Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add some random prefix to split ids. #4824

Open
fulmicoton opened this issue Apr 1, 2024 · 1 comment · May be fixed by #5495
Open

Add some random prefix to split ids. #4824

fulmicoton opened this issue Apr 1, 2024 · 1 comment · May be fixed by #5495
Labels
enhancement New feature or request

Comments

@fulmicoton
Copy link
Contributor

S3 creates partitions based on a prefix of the object keys.
Because we use ULID, splits coming from the same day end up on the same servers.
On the search side, most of the queries tend to target the last few days, and end up hitting the same servers.

For this reason, in some case, we can hit the per-prefix rate limiting.
@fulmicoton fulmicoton added the enhancement New feature or request label Apr 1, 2024
@fulmicoton
Copy link
Contributor Author

https://aws.amazon.com/blogs/aws/amazon-s3-performance-tips-tricks-seattle-hiring-event/

@fulmicoton fulmicoton mentioned this issue Sep 25, 2024
Closed
fulmicoton added a commit that referenced this issue Oct 17, 2024
@fulmicoton
Optionally prepending s3 compatible storage's key with a hash of the …
bb3e77f 
…key.

The point is to workaround S3 rate limiting.
Since it is based on keys, our ULID naming scheme can lead to hotspot in
the keyspace.

This solution has a downside. External scripts listing files will have a their job multiplied.
For this reason, the prefix cardinality is configurable.

Closes #4824
@fulmicoton fulmicoton linked a pull request Oct 17, 2024 that will close this issue
Open
fulmicoton added a commit that referenced this issue Oct 17, 2024
@fulmicoton
Optionally prepending s3 compatible storage's key with a hash of the …
1b0aa58 
…key.

The point is to workaround S3 rate limiting.
Since it is based on keys, our ULID naming scheme can lead to hotspot in
the keyspace.

This solution has a downside. External scripts listing files will have a their job multiplied.
For this reason, the prefix cardinality is configurable.

Closes #4824
fulmicoton added a commit that referenced this issue Oct 17, 2024
@fulmicoton
Optionally prepending s3 compatible storage's key with a hash of the …
d207e72 
…key.

The point is to workaround S3 rate limiting.
Since it is based on keys, our ULID naming scheme can lead to hotspot in
the keyspace.

This solution has a downside. External scripts listing files will have a their job multiplied.
For this reason, the prefix cardinality is configurable.

Closes #4824
fulmicoton added a commit that referenced this issue Oct 17, 2024
@fulmicoton
Optionally prepending s3 compatible storage's key with a hash of the …
8c90035 
…key.

The point is to workaround S3 rate limiting.
Since it is based on keys, our ULID naming scheme can lead to hotspot in
the keyspace.

This solution has a downside. External scripts listing files will have a their job multiplied.
For this reason, the prefix cardinality is configurable.

Closes #4824
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Optionally prepending s3 compatible storage's key with a hash of the … quickwit-oss/quickwit
1 participant
@fulmicoton