Report the running version as a metric

.github/workflows/tests.yml

@@ -87,6 +87,7 @@ jobs:
                 path: ./test-integration.log
                 if-no-files-found: error
     end_to_end:
+        name: E2E tests (maxconns=${{ matrix.max_db_conns }})
         runs-on: ubuntu-latest
         strategy:
           matrix:
@@ -151,14 +152,15 @@ jobs:
                   SYNCV3_DB: user=postgres dbname=syncv3 sslmode=disable password=postgres host=localhost
                   SYNCV3_SERVER: http://localhost:8008
                   SYNCV3_SECRET: itsasecret
+                  SYNCV3_PROM: :2112
                   SYNCV3_MAX_DB_CONN: ${{ matrix.max_db_conns }}
                   E2E_TEST_SERVER_STDOUT: test-e2e-server.log
 
             - name: Upload test log
               uses: actions/upload-artifact@v3
               if: always()
               with:
-                name: E2E test logs
+                name: E2E test logs (maxconns=${{ matrix.max_db_conns }})
                 path: |
                   ./tests-e2e/test-e2e-runner.log
                   ./tests-e2e/test-e2e-server.log

cmd/syncv3/main.go

@@ -16,6 +16,7 @@ import (
 	syncv3 "github.com/matrix-org/sliding-sync"
 	"github.com/matrix-org/sliding-sync/internal"
 	"github.com/matrix-org/sliding-sync/sync2"
+	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	"github.com/rs/zerolog"
 	"go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp"
@@ -117,6 +118,16 @@ func main() {
 				panic(err)
 			}
 		}()
+		// Cribbed from synapse
+		// https://github.com/matrix-org/synapse/blob/d93912042191d30ff1f7aa41d9f0779a609caca8/synapse/metrics/__init__.py#L413-L421
+		// and https://www.robustperception.io/exposing-the-software-version-to-prometheus/
+		versionGauge := prometheus.NewGauge(prometheus.GaugeOpts{
+			Namespace:   "sliding_sync",
+			Name:        "build_info",
+			ConstLabels: map[string]string{"commit": GitCommit, "version": version},
+		})
+		versionGauge.Set(1) // Dummy value; the label is the data here
+		prometheus.MustRegister(versionGauge)
 	}
 	if args[EnvJaeger] != "" {
 		fmt.Printf("Configuring Jaeger collector...\n")

tests-e2e/run-tests.sh

@@ -9,13 +9,30 @@ export SYNCV3_DEBUG=1
 SYNCV3_PID=$!
 trap "kill $SYNCV3_PID" EXIT
 
+waits=0
 # wait for the server to be listening, we want this endpoint to 404 instead of connrefused
 until [ \
   "$(curl -s -w '%{http_code}' -o /dev/null "http://localhost:8844/idonotexist")" \
   -eq 404 ]
 do
   echo 'Waiting for server to start...'
+  waits=$((waits+1))
+  if [ "$waits" -ge 60 ]; then
+    echo "Server did not start within a minute" > /dev/stderr
+    exit 1
+  fi
   sleep 1
 done
 
+# if the metrics env var is set, sanity check that metrics are being served
+if [ -n "${SYNCV3_PROM+1}" ]; then
+  status="$(curl --silent --write-out '%{http_code}' -o /dev/null http://localhost:2112/metrics)"
+  if [ "$status" != 200 ]; then
+    echo "Metrics endpoint returned $status, expected 200" > /dev/stderr
+    exit 1
+  else
+    echo "Metrics endpoint returned 200 OK."
+  fi
+fi
+
 go test "$@"