metrics: count invalid EVENTs, document what nostr_events_total counts (#255)

README.md

@@ -257,7 +257,7 @@ The following metrics are available:
 
 * **`nostr_client_messages_total{verb}`** - Total number of messages received from clients, broken down by verb (EVENT, REQ, CLOSE, NEG-OPEN, NEG-MSG, NEG-CLOSE)
 * **`nostr_relay_messages_total{verb}`** - Total number of messages sent to clients, broken down by verb (EVENT, OK, EOSE, NOTICE, NEG-MSG, NEG-ERR)
-* **`nostr_events_total{kind}`** - Total number of events processed, broken down by event kind (0, 1, 3, 4, etc.)
+* **`nostr_events_total{kind}`** - Total number of events persisted to the DB, broken down by event kind (0, 1, 3, 4, etc.). Rejected and duplicate events are not counted here; see `strfry_write_rejected_total`, `strfry_write_dups_total` and `strfry_invalid_events_total`
 
 To scrape these metrics with Prometheus, add a job to your `prometheus.yml`:
 

src/PrometheusMetrics.h

@@ -101,6 +101,7 @@ class PrometheusMetrics {
     // Write path performance metrics
     Counter writtenEventsTotal;
     Counter rejectedEventsTotal;
+    Counter invalidEventsTotal;  // EVENT messages that failed validation before reaching the writer
     Counter dupEventsTotal;
     Counter writeTimeUs;  // total microseconds spent in write transactions
     Gauge lastWriteBatchSize;
@@ -152,6 +153,10 @@ class PrometheusMetrics {
         out << "# TYPE strfry_write_rejected_total counter\n";
         out << "strfry_write_rejected_total " << rejectedEventsTotal.get() << "\n";
 
+        out << "# HELP strfry_invalid_events_total Total EVENT messages rejected before the write path because they failed validation\n";
+        out << "# TYPE strfry_invalid_events_total counter\n";
+        out << "strfry_invalid_events_total " << invalidEventsTotal.get() << "\n";
+
         out << "# HELP strfry_write_dups_total Total duplicate events skipped\n";
         out << "# TYPE strfry_write_dups_total counter\n";
         out << "strfry_write_dups_total " << dupEventsTotal.get() << "\n";

src/apps/relay/RelayIngester.cpp

@@ -33,6 +33,7 @@ void RelayServer::runIngester(ThreadPool<MsgIngester>::Thread &thr) {
                             try {
                                 ingesterProcessEvent(txn, rsctx, msg->connId, msg->ipAddr, arr[1], writerMsgs);
                             } catch (std::exception &e) {
+                                PrometheusMetrics::getInstance().invalidEventsTotal.inc();
                                 sendOKResponse(msg->connId, arr[1].is_object() && arr[1].at("id").is_string() ? arr[1].at("id").get_string() : "?",
                                                false, std::string("invalid: ") + e.what());
                                 if (cfg().relay__logging__invalidEvents) LI << "Rejected invalid event: " << e.what();