remove unused producer latency metrics

[bmckerry]

We're only using the p99_* producer latency metrics in our dashboards & alerting.
This removes some of our highest-volume metrics.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Autofix prepared a fix for the issue found in the latest run.

• ✅ Fixed: Test call count not updated after metric removal
  • Updated both affected tests to assert the reduced timing call counts that match the remaining p99-only metrics.

Or push these changes by commenting:

@cursor push 1ac364f5d3

Preview (1ac364f5d3)

diff --git a/tests/backends/test_kafka_producer.py b/tests/backends/test_kafka_producer.py
--- a/tests/backends/test_kafka_producer.py
+++ b/tests/backends/test_kafka_producer.py
@@ -24,7 +24,7 @@
 
     producer_stats_callback(stats_json, None)
 
-    assert mock_metrics.timing.call_count == 4
+    assert mock_metrics.timing.call_count == 2
     mock_metrics.timing.assert_any_call(
         "arroyo.producer.librdkafka.p99_int_latency",
         2.0,
@@ -82,7 +82,7 @@
 
     producer_stats_callback(stats_json, None)
 
-    assert mock_metrics.timing.call_count == 6
+    assert mock_metrics.timing.call_count == 3
     mock_metrics.timing.assert_any_call(
         "arroyo.producer.librdkafka.p99_int_latency",
         2.0,

_{This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.}