Experimental caching metrics #3554
Assignees
Labels
Comments
Geal
added a commit
that referenced
this issue
Sep 6, 2023
Fix #3554 This creates a new metric recorded only if we set the configuration option `telemetry.metrics.common.experimental_cache_metrics.enabled` to `true`. * `apollo.router.operations.entity` (histogram): cache hit ratio per subgraph and entity type This simulates an entity cache to find out if it would be useful. Each time we do a subgraph query, we use as cache key: - subgraph name - entity type - query - vary headers - entity key We record if we have seen this entity before (using a bloom filter) and calculate the cache hit ratio for that query, per subgraph and entity type. --------- Signed-off-by: Benjamin Coenen <[email protected]> Co-authored-by: Benjamin Coenen <[email protected]>
garypen
pushed a commit
that referenced
this issue
Sep 12, 2023
Fix #3554 This creates a new metric recorded only if we set the configuration option `telemetry.metrics.common.experimental_cache_metrics.enabled` to `true`. * `apollo.router.operations.entity` (histogram): cache hit ratio per subgraph and entity type This simulates an entity cache to find out if it would be useful. Each time we do a subgraph query, we use as cache key: - subgraph name - entity type - query - vary headers - entity key We record if we have seen this entity before (using a bloom filter) and calculate the cache hit ratio for that query, per subgraph and entity type. --------- Signed-off-by: Benjamin Coenen <[email protected]> Co-authored-by: Benjamin Coenen <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Follow up to #3532 that was reverted. We want metrics to evaluate how much data is reused between queries
The text was updated successfully, but these errors were encountered: