Dump connection summary in TSV format

reports/connections

@@ -0,0 +1,52 @@
+#!/usr/bin/python3
+
+import argparse
+import re
+
+ap = argparse.ArgumentParser()
+ap.add_argument("files", nargs="+")
+
+args = ap.parse_args()
+
+sessions = {}
+summary = {}
+for fn in args.files:
+    with open(fn) as fh:
+        for ln in fh:
+            if m := re.match(r".*user=\[unknown\],db=\[unknown\],pid=(\d+) LOG:  connection received: host=(\S+)", ln):
+                pid = int(m.group(1))
+                host = m.group(2)
+                sessions[pid] = { "host": host }
+            elif m := re.match(r".*user=(.*?),db=(.*?),pid=(\d+) LOG:  connection authorized:", ln):
+                user = m.group(1)
+                db = m.group(2)
+                pid = int(m.group(3))
+                sessions[pid]["user"] = user
+                sessions[pid]["db"] = db
+                if m := re.match(r".*application_name=(\S+)", ln):
+                    sessions[pid]["application_name"] = m.group(1)
+                else:
+                    sessions[pid]["application_name"] = "-"
+            elif m := re.match(r".*user=(.*?),db=(.*?),pid=(\d+) LOG:  disconnection: session time: ([0-9:.]+) user=\S+ database=\S+ host=(\S+)", ln):
+                user = m.group(1)
+                db = m.group(2)
+                pid = int(m.group(3))
+                time_str = m.group(4)
+                host = m.group(5)
+                m = re.match("(\d+):(\d{2}):(\d{2})\.(\d{3})", time_str)
+                time = int(m.group(1)) * 3600 + int(m.group(2)) * 60 + int(m.group(3)) * 1 + int(m.group(4)) * 0.001
+
+                assert user == sessions[pid]["user"]
+                assert db == sessions[pid]["db"]
+                assert host == sessions[pid]["host"]
+
+                s = summary.setdefault((db, user, host, sessions[pid]["application_name"]), { "n_sessions": 0, "session_time": 0 })
+                s["n_sessions"] += 1
+                s["session_time"] += time
+                del sessions[pid]
+
+print("Database", "User", "Client", "Application", "#Sessions", "Session time (s)", sep="\t")
+for k, v in summary.items():
+    print(k[0], k[1], k[2], k[3], v["n_sessions"], v["session_time"], sep="\t")
+
+# vim: tw=0