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

Non-token aware queries graph (and gauge) is broken. #2259

Closed
michoecho opened this issue Apr 5, 2024 · 0 comments · Fixed by #2264
Closed

Non-token aware queries graph (and gauge) is broken. #2259

michoecho opened this issue Apr 5, 2024 · 0 comments · Fixed by #2264
Labels
bug Something isn't working right

Comments

@michoecho
Copy link
Contributor

Panel Name: Non-Token Aware Queries
Dashboard Name: CQL
Scylla-Monitoring Version: master
Scylla-Version: master

There are two problems with this graph:

  1. The combination of Remove counters and LWT from non-token aware #1304 and set_skip_when_empty is broken.
    With set_skip_when_empty, the used scylla_storage_proxy_replica_received_counter_updates metric isn't equal 0, but doesn't exist at all. So the entire formula evaluates to nothing and the entire graph is empty, even if non-token aware queries are actually present.
  2. The formula (scylla_cql_reads - scylla_storage_proxy_coordinator_reads_local_node) is wrong. A coordinator might choose not to serve a replica read even if it belongs to the replica set. (This happens due to heat weighted load balancing — even if the query is token-aware, the coordinator prefers other replicas while its cache is still cold).

But Scylla has metrics which track non-token-aware queries directly. See scylladb/scylladb#4338.

So instead, consider using this for the graph:

rate(scylla_storage_proxy_coordinator_reads_coordinator_outside_replica_set[...]) + rate(scylla_storage_proxy_coordinator_writes_coordinator_outside_replica_set[...])

(with whatever change is needed to make it work even if one of these doesn't exist due to set_skip_when_empty).

@michoecho michoecho added the bug Something isn't working right label Apr 5, 2024
@amnonh amnonh added this to the Monitoring 4.7.1 milestone Apr 14, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working right
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants