test: adapt to changed mz_cluster_replica_metrics

This commit adapts the tests to the changes made to the `mz_cluster_replica_metrics` system relation: * It's no longer a table but a source. * It has an extra column with the event time. * It is a log rather than a snapshot of the most recent state. The `statement-logging` cluster test runs into #29237 when querying the new `mz_cluster_replica_metrics` source, so that query has been disabled until the bug is fixed.
MaterializeInc · Aug 27, 2024 · 351982c · 351982c
1 parent 7e236ec
commit 351982c
Show file tree

Hide file tree

Showing 9 changed files with 25 additions and 18 deletions.
diff --git a/src/environmentd/tests/sql.rs b/src/environmentd/tests/sql.rs
@@ -1372,7 +1372,7 @@ fn test_transactional_explain_timestamps() {
     assert_eq!(*explain_timestamp, mz_now_timestamp);
 }
 
-// Test that the since for `mz_cluster_replica_utilization` is held back by at least
+// Test that the since for `mz_cluster_replicas` is held back by at least
 // 30 days, which is required for the frontend observability work.
 //
 // Feel free to modify this test if that product requirement changes,
@@ -1384,7 +1384,7 @@ async fn test_utilization_hold() {
     // `mz_catalog_server` tests indexes, `quickstart` tests tables.
     // The bool determines whether we are testing indexes.
     const CLUSTERS_TO_TRY: &[&str] = &["mz_catalog_server", "quickstart"];
-    const QUERIES_TO_TRY: &[&str] = &["SELECT * FROM mz_internal.mz_cluster_replica_statuses"];
+    const QUERIES_TO_TRY: &[&str] = &["SELECT * FROM mz_internal.mz_cluster_replicas"];
 
     let now_millis = 619388520000;
     let past_millis = now_millis - THIRTY_DAYS_MS;

diff --git a/test/cloudtest/test_metrics.py b/test/cloudtest/test_metrics.py
@@ -21,8 +21,17 @@ def test_replica_metrics(mz: MaterializeApplication) -> None:
             """
             > CREATE CLUSTER my_cluster REPLICAS (my_replica (SIZE '4-4'))
 
-            > SELECT COUNT(*) FROM mz_internal.mz_cluster_replica_metrics m JOIN mz_cluster_replicas cr ON m.replica_id = cr.id WHERE cr.name = 'my_replica' AND m.cpu_nano_cores IS NOT NULL AND m.memory_bytes IS NOT NULL
-            4
+            > SELECT DISTINCT process_id
+              FROM mz_internal.mz_cluster_replica_metrics m
+              JOIN mz_cluster_replicas cr ON m.replica_id = cr.id
+              WHERE
+                  cr.name = 'my_replica' AND
+                  m.cpu_nano_cores IS NOT NULL AND
+                  m.memory_bytes IS NOT NULL
+            0
+            1
+            2
+            3
 
             > DROP CLUSTER my_cluster
             """

diff --git a/test/cluster/statement-logging/statement-logging.td b/test/cluster/statement-logging/statement-logging.td
@@ -94,8 +94,9 @@ contains: column "f_nonexistent" does not exist
 > DROP CLUSTER c
 
 # This will be executed on mz_catalog_server, due to auto-routing of catalog queries.
-> SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics
-true
+# Disabled until #29237 is fixed, as the query lifecycle check below is flaky otherwise.
+#> SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics
+#true
 
 # Test that we're correctly recording transaction isolation
 > SET transaction_isolation TO serializable
@@ -145,7 +146,6 @@ quickstart quickstart my_app 1 {} success <null> <null> <null> true "PREPARE p A
 quickstart quickstart my_app 1 {} success <null> <null> <null> true "SET cluster TO c" "strict serializable" <null> materialize {public}
 mz_catalog_server mz_catalog_server my_app 1 {} success <null> 1 constant true "SELECT 'beginning real test!'" "strict serializable" <null> materialize {public}
 mz_catalog_server mz_catalog_server my_app 1 {} success <null> 1 constant true "SELECT 'serializable'" serializable <null> materialize {public}
-mz_catalog_server mz_catalog_server my_app 1 {} success <null> 1 standard true "SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics" "strict serializable" true materialize {public}
 
 > WITH all_stmts AS (SELECT mseh.id, mseh.began_at, mst.sql FROM mz_internal.mz_statement_execution_history mseh JOIN mz_internal.mz_prepared_statement_history mpsh ON mseh.prepared_statement_id = mpsh.id JOIN (SELECT DISTINCT sql, sql_hash, redacted_sql FROM mz_internal.mz_sql_text) mst ON mpsh.sql_hash = mst.sql_hash),
        test_begin AS (SELECT began_at FROM all_stmts WHERE sql = 'SELECT ''beginning real test!''' ORDER BY began_at DESC LIMIT 1)
@@ -199,11 +199,6 @@ COMMIT execution-finished
 "SELECT * FROM t" execution-finished
 "SELECT * FROM t" optimization-finished
 "SELECT * FROM t" storage-dependencies-finished
-"SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics" compute-dependencies-finished
-"SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics" execution-began
-"SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics" execution-finished
-"SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics" optimization-finished
-"SELECT count(*) > 0 FROM mz_internal.mz_cluster_replica_metrics" storage-dependencies-finished
 "SELECT count(*) FROM t" compute-dependencies-finished
 "SELECT count(*) FROM t" compute-dependencies-finished
 "SELECT count(*) FROM t" execution-began

diff --git a/test/sqllogictest/autogenerated/mz_internal.slt b/test/sqllogictest/autogenerated/mz_internal.slt
@@ -116,6 +116,7 @@ SELECT position, name, type FROM objects WHERE schema = 'mz_internal' AND object
 3  cpu_nano_cores  uint8
 4  memory_bytes  uint8
 5  disk_bytes  uint8
+6  updated_at  timestamp␠with␠time␠zone
 
 query ITT
 SELECT position, name, type FROM objects WHERE schema = 'mz_internal' AND object = 'mz_cluster_replica_statuses' ORDER BY position
@@ -134,6 +135,7 @@ SELECT position, name, type FROM objects WHERE schema = 'mz_internal' AND object
 3  cpu_percent  double␠precision
 4  memory_percent  double␠precision
 5  disk_percent  double␠precision
+6  updated_at  timestamp␠with␠time␠zone
 
 query ITT
 SELECT position, name, type FROM objects WHERE schema = 'mz_internal' AND object = 'mz_cluster_replica_history' ORDER BY position

diff --git a/test/sqllogictest/information_schema_tables.slt b/test/sqllogictest/information_schema_tables.slt
@@ -282,7 +282,7 @@ VIEW
 materialize
 mz_internal
 mz_cluster_replica_metrics
-BASE TABLE
+SOURCE
 materialize
 mz_internal
 mz_cluster_replica_statuses

diff --git a/test/sqllogictest/mz_catalog_server_index_accounting.slt b/test/sqllogictest/mz_catalog_server_index_accounting.slt
@@ -208,6 +208,7 @@ mz_cluster_replica_metrics  disk_bytes
 mz_cluster_replica_metrics  memory_bytes
 mz_cluster_replica_metrics  process_id
 mz_cluster_replica_metrics  replica_id
+mz_cluster_replica_metrics  updated_at
 mz_cluster_replica_sizes  cpu_nano_cores
 mz_cluster_replica_sizes  credits_per_hour
 mz_cluster_replica_sizes  disk_bytes

diff --git a/test/sqllogictest/system-cluster.slt b/test/sqllogictest/system-cluster.slt
@@ -410,7 +410,7 @@ ORDER BY r.id;
 ----
 Explained Query:
   Finish order_by=[#0 asc nulls_last] output=[#0..=#5]
-    Project (#0..=#3, #5, #29)
+    Project (#0..=#3, #5, #30)
       Map (((uint8_to_double(#27) / uint8_to_double(#21)) * 100))
         Join on=(#0 = #15 = #24 AND #2 = #4 AND #16 = #17) type=delta
           ArrangeBy keys=[[#0], [#2]]

diff --git a/test/testdrive/catalog.td b/test/testdrive/catalog.td
@@ -565,6 +565,7 @@ name                                         type    cluster    comment
 -----------------------------------------------------------------------
 mz_aws_privatelink_connection_status_history source  <null>     ""
 mz_cluster_replica_frontiers                 source  <null>     ""
+mz_cluster_replica_metrics                   source  <null>     ""
 mz_compute_dependencies                      source  <null>     ""
 mz_compute_error_counts_raw_unified          source  <null>     ""
 mz_compute_hydration_times                   source  <null>     ""
@@ -587,7 +588,6 @@ name                                     comment
 ------------------------------------------------
 mz_aggregates                            ""
 mz_aws_connections                       ""
-mz_cluster_replica_metrics               ""
 mz_cluster_replica_statuses              ""
 mz_cluster_schedules                     ""
 mz_cluster_workload_classes              ""
@@ -769,7 +769,7 @@ test_table ""
 
 # `SHOW TABLES` and `mz_tables` should agree.
 > SELECT COUNT(*) FROM mz_tables WHERE id LIKE 's%'
-56
+55
 
 # There is one entry in mz_indexes for each field_number/expression of the index.
 > SELECT COUNT(id) FROM mz_indexes WHERE id LIKE 's%'

diff --git a/test/testdrive/metrics-cluster.td b/test/testdrive/metrics-cluster.td
@@ -11,7 +11,7 @@
 > CREATE CLUSTER bar REPLICAS ()
 > CREATE CLUSTER xyzzy REPLICAS (size_4 (SIZE '4'))
 
-> SELECT
+> SELECT DISTINCT ON (c.name, r.name, u.process_id)
       c.name,
       r.name,
       u.process_id,
@@ -25,7 +25,7 @@
               mz_internal.mz_cluster_replica_utilization AS u
               ON r.id = u.replica_id
   WHERE c.name IN ( 'foo', 'bar', 'xyzzy' )
-  ORDER BY c.name, r.name, process_id
+  ORDER BY c.name, r.name, process_id, updated_at DESC
 foo size_1 0 true true <null>
 foo size_2_2 0 true true <null>
 foo size_2_2 1 true true <null>