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.

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

GCP: Don't try to read tiles which very probably don't exist #69

Merged

Conversation

AlCutter
Copy link
Collaborator

@AlCutter AlCutter commented Jul 17, 2024

This PR detects when the cache is requesting tiles which are beyond the extent of the current tree and skips the read.

These "future" tiles are very likely not to exist, and, since they're requested by the visitor, we end up trying to fetching serially which adds a fair chunk of latency to an integration run. Not fetching these future tiles speeds the integration up by around 25%.

While it's possible that these tiles could occasionally exist (e.g. due to an integration run crashing previously), not reading should still be safe as the storage design makes integration an idempotent operation, and this is enforced when tiles are finally flushed to GCS at the end of the integration run.

Toward #23

@AlCutter AlCutter force-pushed the prewarm_probably_empty_tiles branch from fed9adb to 91000ae Compare July 17, 2024 15:15
@AlCutter AlCutter merged commit 64a40c2 into transparency-dev:main Jul 17, 2024
4 checks passed
@AlCutter AlCutter deleted the prewarm_probably_empty_tiles branch July 17, 2024 15:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants