sheetsync: Record counts of rows in the DB, segmented by various columns

This lets us view a number of useful graphs in dashboards, eg. rows by state,
errored rows, rows by day, rows by category, meltdowns per day, fraction of
events that are poster moments by category.

Sheetsync was the natural place to do this since it was already periodically scanning
the entire events table.

monitoring/dashboards/overview.jsonnet

@@ -151,8 +151,12 @@ grafana.dashboard({
           {
             name: "Database events by state",
             axis: {min: 0, label: "events"},
-            tooltip: "Not implemented", // TODO
+            stack: true,
-            expressions: {"Not implemented": "0"},
+            expressions: {
+              "{{state}}": |||
+                sum(event_counts) by (state)
+              |||,
+            },
           },
         ],
       ],

sheetsync/sheetsync/main.py

@@ -3,6 +3,7 @@ import json
 import logging
 import signal
 import uuid
+from collections import defaultdict
 
 import argh
 import gevent.backdoor
@@ -41,6 +42,12 @@ rows_changed = prom.Counter(
 	['type', 'worksheet'],
 )
 
+event_counts = prom.Gauge(
+	'event_counts',
+	'Number of rows in the database',
+	['sheet_name', 'category', 'poster_moment', 'state', 'errored'],
+)
+
 class SheetSync(object):
 
 	# Time between syncs
@@ -190,8 +197,12 @@ class SheetSync(object):
 		"""Return the entire events table as a map {id: event namedtuple}"""
 		result = query(self.conn, "SELECT * FROM events")
 		by_id = {}
+		counts = defaultdict(lambda: 0)
 		for row in result.fetchall():
 			by_id[row.id] = row
+			counts[row.sheet_name, row.category, str(row.poster_moment), row.state, str(bool(row.error))] += 1
+		for labels, count in counts.items():
+			event_counts.labels(*labels).set(count)
 		return by_id
 
 	def parse_row(self, row):