[tmpnet] Misc cleanup for monitoring tooling (#3527)

scripts/configure-local-metrics-collection.sh

@@ -51,4 +51,9 @@ echo "Wrote promtail configuration to ${PROMTAIL_CONFIG_FILE}"
 
 echo "Metrics collection by prometheus can be started with ./scripts/run_prometheus.sh"
 echo "Log collection by promtail can be started with ./scripts/run_promtail.sh"
-echo "Grafana link: https://grafana-poc.avax-dev.network/d/kBQpRdWnk/avalanche-main-dashboard?var-filter=network_uuid%7C%3D%7C${METRICS_UUID}"
+
+GRAFANA_LINK="https://grafana-poc.avax-dev.network/d/kBQpRdWnk/avalanche-main-dashboard?var-filter=network_uuid%7C%3D%7C${METRICS_UUID}"
+METRICS_PATH="${HOME}/.avalanchego/metrics.txt"
+echo "${GRAFANA_LINK}" > "${METRICS_PATH}"
+echo "Metrics and logs can be viewed at: ${GRAFANA_LINK}"
+echo "Link also saved to ${METRICS_PATH}"

scripts/run_prometheus.sh

@@ -2,17 +2,16 @@
 
 set -euo pipefail
 
-# Starts a prometheus instance in agent-mode, forwarding to a central
-# instance. Intended to enable metrics collection from temporary networks running
-# locally and in CI.
+# - Starts a prometheus instance in agent-mode to collect metrics from nodes running
+#   locally and in CI.
 #
-# The prometheus instance will remain running in the background and will forward
-# metrics to the central instance for all tmpnet networks.
+# - promtail will remain running in the background and will forward metrics to the
+#   specified prometheus endpoint.
 #
-# To stop it:
-#
-#   $ kill -9 `cat ~/.tmpnet/prometheus/run.pid` && rm ~/.tmpnet/prometheus/run.pid
+# - Each node is configured with a file written to ~/.tmpnet/prometheus/file_sd_configs
 #
+# - To stop the running instance:
+#     $ kill -9 `cat ~/.tmpnet/promtheus/run.pid` && rm ~/.tmpnet/promtail/run.pid
 
 # e.g.,
 # PROMETHEUS_ID=<id> PROMETHEUS_PASSWORD=<password> ./scripts/run_prometheus.sh
@@ -45,7 +44,7 @@ fi
 
 PROMETHEUS_PASSWORD="${PROMETHEUS_PASSWORD:-}"
 if [[ -z "${PROMETHEUS_PASSWORD}" ]]; then
-  echo "Plase provide a value for PROMETHEUS_PASSWORD"
+  echo "Please provide a value for PROMETHEUS_PASSWORD"
   exit 1
 fi
 
@@ -64,13 +63,13 @@ if ! command -v "${CMD}" &> /dev/null; then
 
     # Determine the arch
     if which sw_vers &> /dev/null; then
-      echo "on macos, only amd64 binaries are available so rosetta is required on apple silicon machines."
-      echo "to avoid using rosetta, install via homebrew: brew install prometheus"
+      echo "On macos, only amd64 binaries are available so rosetta is required on apple silicon machines."
+      echo "To avoid using rosetta, install via homebrew: brew install prometheus"
       DIST=darwin
     else
       ARCH="$(uname -i)"
       if [[ "${ARCH}" != "x86_64" ]]; then
-        echo "on linux, only amd64 binaries are available. manual installation of prometheus is required."
+        echo "On linux, only amd64 binaries are available. manual installation of prometheus is required."
         exit 1
       else
         DIST="linux"
@@ -90,8 +89,8 @@ fi
 FILE_SD_PATH="${PROMETHEUS_WORKING_DIR}/file_sd_configs"
 mkdir -p "${FILE_SD_PATH}"
 
-echo "writing configuration..."
-cat >"${PROMETHEUS_WORKING_DIR}"/prometheus.yaml <<EOL
+CONFIG_PATH="${PROMETHEUS_WORKING_DIR}/prometheus.yaml"
+cat > "${CONFIG_PATH}" <<EOL
 # my global config
 global:
   # Make sure this value takes into account the network-shutdown-delay in tests/fixture/e2e/env.go
@@ -112,9 +111,12 @@ remote_write:
       username: "${PROMETHEUS_ID}"
       password: "${PROMETHEUS_PASSWORD}"
 EOL
+echo "Wrote configuration to ${CONFIG_PATH}"
 
-echo "starting prometheus..."
+echo "Starting prometheus..."
 cd "${PROMETHEUS_WORKING_DIR}"
 nohup "${CMD}" --config.file=prometheus.yaml --web.listen-address=localhost:0 --enable-feature=agent > prometheus.log 2>&1 &
 echo $! > "${PIDFILE}"
-echo "running with pid $(cat "${PIDFILE}")"
+echo "prometheus started with pid $(cat "${PIDFILE}")"
+# shellcheck disable=SC2016
+echo 'To stop prometheus: "kill -SIGTERM `cat ~/.tmpnet/prometheus/run.pid` && rm ~/.tmpnet/prometheus/run.pid"'

scripts/run_promtail.sh

@@ -2,16 +2,15 @@
 
 set -euo pipefail
 
-# Starts a promtail instance to collect logs from temporary networks
-# running locally and in CI.
+# - Starts a promtail instance to collect logs from nodes running locally and in CI.
 #
-# The promtail instance will remain running in the background and will forward
-# logs to the central instance for all tmpnet networks.
+# - promtail will remain running in the background and will forward logs to the
+#   specified Loki endpoint.
 #
-# To stop it:
-#
-#   $ kill -9 `cat ~/.tmpnet/promtail/run.pid` && rm ~/.tmpnet/promtail/run.pid
+# - Each node is configured with a file written to ~/.tmpnet/promtail/file_sd_configs/
 #
+# - To stop the running instance:
+#     $ kill -9 `cat ~/.tmpnet/promtail/run.pid` && rm ~/.tmpnet/promtail/run.pid
 
 # e.g.,
 # LOKI_ID=<id> LOKI_PASSWORD=<password> ./scripts/run_promtail.sh
@@ -44,7 +43,7 @@ fi
 
 LOKI_PASSWORD="${LOKI_PASSWORD:-}"
 if [[ -z "${LOKI_PASSWORD}" ]]; then
-  echo "Plase provide a value for LOKI_PASSWORD"
+  echo "Please provide a value for LOKI_PASSWORD"
   exit 1
 fi
 
@@ -86,8 +85,8 @@ fi
 FILE_SD_PATH="${PROMTAIL_WORKING_DIR}/file_sd_configs"
 mkdir -p "${FILE_SD_PATH}"
 
-echo "writing configuration..."
-cat >"${PROMTAIL_WORKING_DIR}"/promtail.yaml <<EOL
+CONFIG_PATH="${PROMTAIL_WORKING_DIR}/promtail.yaml"
+cat > "${CONFIG_PATH}" <<EOL
 server:
   http_listen_port: 0
   grpc_listen_port: 0
@@ -107,9 +106,12 @@ scrape_configs:
       - files:
           - '${FILE_SD_PATH}/*.json'
 EOL
+echo "Wrote configuration to ${CONFIG_PATH}"
 
-echo "starting promtail..."
+echo "Starting promtail..."
 cd "${PROMTAIL_WORKING_DIR}"
 nohup "${CMD}" -config.file=promtail.yaml > promtail.log 2>&1 &
 echo $! > "${PIDFILE}"
-echo "running with pid $(cat "${PIDFILE}")"
+echo "promtail started with pid $(cat "${PIDFILE}")"
+# shellcheck disable=SC2016
+echo 'To stop promtail: "kill -SIGTERM `cat ~/.tmpnet/promtail/run.pid` && rm ~/.tmpnet/promtail/run.pid"'

tests/fixture/tmpnet/README.md

@@ -167,6 +167,7 @@ HOME
             │       └── config.json                      // Custom chain configuration for all nodes
             ├── config.json                              // Common configuration (including defaults and pre-funded keys)
             ├── genesis.json                             // Genesis for all nodes
+            ├── metrics.txt                              // Link for metrics and logs collected from the network (see: Monitoring)
             ├── network.env                              // Sets network dir env var to simplify network usage
             └── subnets                                  // Directory containing subnet config for both avalanchego and tmpnet
                 ├── subnet-a.json                        // tmpnet configuration for subnet-a and its chain(s)
@@ -269,6 +270,11 @@ LOKI_ID=<id> LOKI_PASSWORD=<password> ./scripts/run_promtail.sh
 
 # Network start emits link to grafana displaying collected logs and metrics
 ./build/tmpnetctl start-network
+
+# Configure metrics collection from a local node binding to the default API
+# port of 9650 and storing its logs in ~/.avalanchego/logs. The script will
+# also emit a link to grafana.
+./scripts/configure-local-metrics-collection.sh
 ```
 
 ### Metrics collection

tests/fixture/tmpnet/network.go

@@ -373,10 +373,17 @@ func (n *Network) StartNodes(ctx context.Context, w io.Writer, nodesToStart ...*
 	if _, err := fmt.Fprintf(w, "\nStarted network %s (UUID: %s)\n", n.Dir, n.UUID); err != nil {
 		return err
 	}
-	// Provide a link to the main dashboard filtered by the uuid and showing results from now till whenever the link is viewed
+
+	// Generate a link to the main dashboard filtered by the uuid and showing results from now till whenever the link is viewed
 	startTimeStr := strconv.FormatInt(startTime.UnixMilli(), 10)
 	metricsURL := MetricsLinkForNetwork(n.UUID, startTimeStr, "")
-	if _, err := fmt.Fprintf(w, "\nMetrics: %s\n", metricsURL); err != nil {
+
+	// Write link to the network path and to stdout
+	metricsPath := filepath.Join(n.Dir, "metrics.txt")
+	if err := os.WriteFile(metricsPath, []byte(metricsURL+"\n"), perms.ReadWrite); err != nil {
+		return fmt.Errorf("failed to write metrics link to %s: %w", metricsPath, err)
+	}
+	if _, err := fmt.Fprintf(w, "\nMetrics: %s\nLink also saved to %s\n", metricsURL, metricsPath); err != nil {
 		return err
 	}