measure usage

_sql/analytics/generic_inference/jobs_by_ip_address.sql

@@ -6,7 +6,8 @@
 -- NB: Nevermind, there's no index on `created_at`
 -- select count(*) from generic_inference_jobs where created_at > now() - interval 1 minute;
 
--- When the database is under contention, use this form to reduce the number of rows scanned
+-- Bin histogram by (creator user, IP address) (including logged out)
+-- NB: no index on ip, hence subquery
 SELECT
     maybe_creator_user_token,
     users.username as maybe_creator_username,
@@ -32,32 +33,11 @@ group by maybe_creator_user_token, maybe_creator_username, creator_ip_address
 order by attempts desc;
 
 
--- When the database is under contention, use this form to reduce the number of rows scanned
+-- Bin histogram by creator user (including logged out)
+-- NB: no index on ip, hence subquery
 SELECT
-    creator_ip_address,
-    count(*) as attempts
-FROM (
-    SELECT
-        maybe_creator_user_token,
-        creator_ip_address
-    FROM (
-        SELECT maybe_creator_user_token,
-            creator_ip_address,
-            created_at
-        FROM generic_inference_jobs
-        ORDER BY id DESC
-        LIMIT 10000
-    ) as j
-    WHERE j.created_at > NOW() - INTERVAL 30 MINUTE
-) as jobs
-group by creator_ip_address
-order by attempts desc;
-
-
--- When the database is under contention, use this form to reduce the number of rows scanned
-SELECT
-    maybe_creator_user_token,
     users.username as maybe_creator_username,
+    maybe_creator_user_token,
     count(*) as attempts
 FROM (
     SELECT
@@ -77,3 +57,50 @@ LEFT OUTER JOIN users
 ON users.token = jobs.maybe_creator_user_token
 group by maybe_creator_user_token, maybe_creator_username
 order by attempts desc;
+
+
+-- Bin histogram by IP (logged in and non-logged in
+-- NB: no index on ip, hence subquery
+SELECT
+    creator_ip_address,
+    count(*) as attempts
+FROM (
+     SELECT
+         maybe_creator_user_token,
+         creator_ip_address
+     FROM (
+          SELECT maybe_creator_user_token,
+                 creator_ip_address,
+                 created_at
+          FROM generic_inference_jobs
+          ORDER BY id DESC
+              LIMIT 10000
+      ) as j
+     WHERE j.created_at > NOW() - INTERVAL 30 MINUTE
+ ) as jobs
+group by creator_ip_address
+order by attempts desc;
+
+
+-- Bin histogram by non-logged in IP
+-- NB: no index on ip, hence subquery
+SELECT
+    creator_ip_address,
+    count(*) as attempts
+FROM (
+     SELECT
+         maybe_creator_user_token,
+         creator_ip_address
+     FROM (
+          SELECT maybe_creator_user_token,
+                 creator_ip_address,
+                 created_at
+          FROM generic_inference_jobs
+          ORDER BY id DESC
+              LIMIT 10000
+      ) as j
+     WHERE j.created_at > NOW() - INTERVAL 30 MINUTE
+     AND j.maybe_creator_user_token IS NULL
+ ) as jobs
+group by creator_ip_address
+order by attempts desc;