Update Grafana

Update grafana, to use more recent golang. This also brings support of multi line logs in loki.

See merge request nexedi/slapos!921

component/nodejs/buildout.cfg

@@ -14,6 +14,12 @@ parts =
 [nodejs]
 <= nodejs-8.9.4
 
+[nodejs-14.16.0]
+<= nodejs-base
+openssl_location = ${openssl:location}
+version = v14.16.0
+md5sum = 7dc3666f407bf4e12a01ce1be2883d31
+
 [nodejs-12.18.3]
 <= nodejs-base
 openssl_location = ${openssl:location}

software/grafana/README.md

@@ -8,10 +8,10 @@ This is an experimental integration, mainly to evaluate these solutions.
 See https://github.com/influxdata/telegraf to learn about plugins.
 
 Useful plugins in this context are probably
-[exec](https://github.com/influxdata/telegraf/tree/1.11.1/plugins/inputs/exec),
-[logparser](https://github.com/influxdata/telegraf/tree/1.11.1/plugins/inputs/logparser)
+[exec](https://github.com/influxdata/telegraf/tree/v1.17.3/plugins/inputs/exec),
+[logparser](https://github.com/influxdata/telegraf/tree/v1.17.3/plugins/inputs/logparser)
 or
-[http](https://github.com/influxdata/telegraf/tree/1.11.1/plugins/inputs/http).
+[http](https://github.com/influxdata/telegraf/tree/v1.17.3/plugins/inputs/http).
 
 Telegraf will save in the `telegraf` database from the embedded influxdb server.
 

software/grafana/buildout.hash.cfg

@@ -15,7 +15,7 @@
 
 [instance-profile]
 filename = instance.cfg.in
-md5sum = 13f4f3806522b265dd1912ff169d146d
+md5sum = 3ccdd2299e759488545b62368c7a0b91
 
 [influxdb-config-file]
 filename = influxdb-config-file.cfg.in
@@ -35,8 +35,8 @@ md5sum = 3aa0f1ed752b2a59ea2b5e7c1733daf3
 
 [loki-config-file]
 filename = loki-config-file.cfg.in
-md5sum = 02ba5acf23fcf88f5594919f46838533
+md5sum = ad2baf4599a937d7352034a41fa24814
 
 [promtail-config-file]
 filename = promtail-config-file.cfg.in
-md5sum = c77788d0a3cc654ad9393eb4b1f31e94
+md5sum = c8c9d815dd7b427788c066f041f04573

software/grafana/gowork.cfg

@@ -20,22 +20,22 @@ revision      = 1f7c19e5f5
 <= go-git-package
 go.importpath = github.com/grafana/grafana
 repository    = https://github.com/grafana/grafana
-revision      = v6.3.0-0-g830da0fda0
+revision      = v7.4.3-0-g010f20c1c8
 
 [go_github.com_grafana_loki]
 <= go-git-package
 go.importpath = github.com/grafana/loki
 repository    = https://github.com/grafana/loki
-revision      = v0.3.0-0-gc673380bb3
+revision      = v2.1.0-188-g877f524c36
 
 [go_github.com_influxdata_influxdb]
 <= go-git-package
 go.importpath = github.com/influxdata/influxdb
 repository    = https://github.com/influxdata/influxdb
-revision      = v1.7.8rc1-0-ga7afa2ecf1
+revision      = v1.8.4-0-gbc8ec4384e
 
 [go_github.com_influxdata_telegraf]
 <= go-git-package
 go.importpath = github.com/influxdata/telegraf
 repository    = https://github.com/influxdata/telegraf
-revision      = 1.11.1-0-gf91de60d37
+revision      = v1.17.3-0-g24a552b90b

software/grafana/instance.cfg.in

@@ -132,7 +132,7 @@ command-line =
   bash -c "{{ influx_bin }} \
      -username ${influxdb:auth-username} \
      -password ${influxdb:auth-password} \
-     -host ${influxdb:host} \
+     -host [${influxdb:host}] \
      -port ${influxdb:http-port} \
      -unsafeSsl \
      -ssl \
@@ -219,6 +219,7 @@ storage-boltdb-dir = ${directory:loki-storage-boltdb-dir}
 storage-filesystem-dir = ${directory:loki-storage-filesystem-dir}
 ip = ${instance-parameter:ipv4-random}
 port = 3100
+grpc-port = 9095
 url = http://${:ip}:${:port}
 
 
@@ -238,9 +239,10 @@ command-line =
 wrapper-path = ${directory:service}/promtail
 
 dir = ${directory:promtail-dir}
-http_port = 19080
+http-port = 19080
+grpc-port = 19095
 ip = ${instance-parameter:ipv4-random}
-url = http://${:ip}:${:http_port}
+url = http://${:ip}:${:http-port}
 
 [promtail-config-file]
 <= config-file
@@ -252,7 +254,7 @@ context =
 [promtail-listen-promise]
 <= check-port-listening-promise
 hostname= ${promtail:ip}
-port = ${promtail:http_port}
+port = ${promtail:http-port}
 
 
 

software/grafana/loki-config-file.cfg.in

 auth_enabled: false
 
 server:
+  http_listen_address: {{ loki['ip'] }}
   http_listen_port: {{ loki['port'] }}
+  grpc_listen_address: {{ loki['ip'] }}
+  grpc_listen_port: {{ loki['grpc-port'] }}
 
 ingester:
   lifecycler:

software/grafana/promtail-config-file.cfg.in

-# https://github.com/grafana/loki/blob/master/docs/logentry/processing-log-lines.md
 
 server:
-  http_listen_port: {{ promtail['http_port'] }}
-  # XXX this external_url does not work ... promtail still listen on all IPs
+  http_listen_address: {{ promtail['ip'] }}
+  http_listen_port: {{ promtail['http-port'] }}
+  grpc_listen_address: {{ promtail['ip'] }}
+  grpc_listen_port: {{ promtail['grpc-port'] }}
   external_url: {{ promtail['url'] }}
-  grpc_listen_port: 0
 
 positions:
   filename: {{ promtail['dir'] }}/positions.yaml

software/grafana/software.cfg

@@ -22,12 +22,14 @@ parts =
   loki-config-file
   promtail-config-file
 
+[python]
+part = python3
 
 [nodejs]
-<= nodejs-10.6.0
+<= nodejs-14.16.0
 
 [yarn]
-<= yarn-1.16.0
+<= yarn-1.22.10
 
 [gowork]
 # All the softwares installed in the go workspace have "non standard" installation
@@ -44,7 +46,7 @@ loki-bin = ${:bin}/loki
 promtail-bin = ${:bin}/promtail
 
 # use recent go
-golang  = ${golang1.12:location}
+golang  = ${golang1.16:location}
 
 [gowork.goinstall]
 command = :
@@ -60,8 +62,7 @@ depends =
 command = bash -c ". ${gowork:env.sh} && \
    go install -v github.com/golang/dep/cmd/dep && \
    cd ${gowork:directory}/src/github.com/influxdata/influxdb && \
-   dep ensure && \
-   go install ./cmd/..."
+   go install ./..."
 update-command =
 
 [telegraf-install]
@@ -75,14 +76,17 @@ update-command =
 
 [grafana-install]
 <= gowork.goinstall
-# yarn and go run build.go needs our nodejs in $PATH
-command = bash -c "export PATH=${nodejs:location}/bin/:$PATH && \
-   . ${gowork:env.sh} && \
+command = bash -c ". ${gowork:env.sh} && \
    cd ${gowork:directory}/src/github.com/grafana/grafana && \
    go run build.go setup && \
    go run build.go build && \
    ${yarn:location}/bin/yarn install --pure-lockfile && \
-   ${nodejs:location}/bin/npm run build"
+   ${yarn:location}/bin/yarn run build && \
+   ${yarn:location}/bin/yarn run plugins:build-bundled && \
+   # Cleanup yarn and Cypress caches
+   rm -rf ${buildout:directory}/.cache/Cypress/ && \
+   rm -rf ${buildout:directory}/.cache/yarn/
+   "
 update-command =
 
 [loki-install]

software/grafana/test/test.py

@@ -25,12 +25,13 @@
 #
 ##############################################################################
 
-import os
-import textwrap
 import logging
+import os
 import tempfile
+import textwrap
 import time
 
+import psutil
 import requests
 
 from slapos.testing.testcase import makeModuleSetUpAndTestCaseClass
@@ -44,11 +45,11 @@ setUpModule, SlapOSInstanceTestCase = makeModuleSetUpAndTestCaseClass(
 class GrafanaTestCase(SlapOSInstanceTestCase):
   """Base test case for grafana.
 
-  Since the instances takes timte to start and stop,
-  we increate as lot the number of retries.
+  Since the instances takes time to start and stop,
+  we increase the number of retries.
   """
+  instance_max_retry = 50
   report_max_retry = 30
-  instance_max_retry = 30
 
 
 class TestGrafana(GrafanaTestCase):
@@ -144,21 +145,27 @@ class TestLoki(GrafanaTestCase):
                 r'''
                 - job_name: {cls.__name__}
                   pipeline_stages:
-                    - regex:
-                        expression: "^(?P<timestamp>.*) - (?P<name>\\S+) - (?P<level>\\S+) - (?P<message>.*)"
-                    - timestamp:
-                        format: 2006-01-02T15:04:05Z00:00
-                        source: timestamp
-                    - labels:
-                        level:
-                        name:
+                    - match:
+                        selector: '{{job="{cls.__name__}"}}'
+                        stages:
+                          - multiline:
+                              firstline: '^\d{{4}}-\d{{2}}-\d{{2}}\s\d{{1,2}}\:\d{{2}}\:\d{{2}}\,\d{{3}}'
+                              max_wait_time: 3s
+                          - regex:
+                              expression: '^(?P<timestamp>.*) - (?P<name>\S+) - (?P<level>\S+) - (?P<message>.*)'
+                          - timestamp:
+                              format: 2006-01-02T15:04:05Z00:00
+                              source: timestamp
+                          - labels:
+                              level:
+                              name:
                   static_configs:
-                  - targets:
-                      - localhost
-                    labels:
-                      job: {cls.__name__}
-                      __path__: {cls._logfile.name}
-                ''').format(**locals())
+                    - targets:
+                        - localhost
+                      labels:
+                        job: {cls.__name__}
+                        __path__: {cls._logfile.name}
+            ''').format(**locals())
     }
 
   @classmethod
@@ -180,6 +187,7 @@ class TestLoki(GrafanaTestCase):
     # create a logger logging to the file that we have
     # configured in instance parameter.
     test_logger = logging.getLogger(self.id())
+    test_logger.propagate = False
     test_logger.setLevel(logging.INFO)
     test_handler = logging.FileHandler(filename=self._logfile.name)
     test_handler.setFormatter(
@@ -189,16 +197,25 @@ class TestLoki(GrafanaTestCase):
     test_logger.info("testing message")
     test_logger.info("testing another message")
     test_logger.warning("testing warn")
+    # log an exception, which will be multi line in log file.
+    def nested1():
+      def nested2():
+        raise ValueError('boom')
+      nested2()
+    try:
+      nested1()
+    except ValueError:
+      test_logger.exception("testing exception")
 
     # Check our messages have been ingested
     # we retry a few times, because there's a short delay until messages are
     # ingested and returned.
-    for i in range(10):
+    for i in range(60):
       resp = requests.get(
           '{self.loki_url}/api/prom/query?query={{job="TestLoki"}}'.format(
               **locals()),
           verify=False).json()
-      if not resp:
+      if len(resp.get('streams', [])) < 3:
         time.sleep(0.5 * i)
         continue
 
@@ -220,9 +237,83 @@ class TestLoki(GrafanaTestCase):
             line for line in info_stream['entries']
             if "testing another message" in line['line']
         ])
+
+    error_stream_list = [stream for stream in resp['streams'] if 'level="ERROR"' in stream['labels']]
+    self.assertEqual(1, len(error_stream_list), resp['streams'])
+    error_stream, = error_stream_list
+    line, = [line['line'] for line in error_stream['entries']]
+    # this entry is multi-line
+    self.assertIn('testing exception\nTraceback (most recent call last):\n', line)
+    self.assertIn('ValueError: boom', line)
+
     # The labels we have configued are also available
     resp = requests.get(
         '{self.loki_url}/api/prom/label'.format(**locals()),
         verify=False).json()
     self.assertIn('level', resp['values'])
     self.assertIn('name', resp['values'])
+
+
+class TestListenInPartition(GrafanaTestCase):
+  def setUp(self):
+    with self.slap.instance_supervisor_rpc as supervisor:
+      all_process_info = supervisor.getAllProcessInfo()
+
+    self.process_dict = {
+        p['name'].replace('-on-watch', ''): psutil.Process(p['pid'])
+        for p in all_process_info if p['name'] != 'watchdog'
+    }
+
+  def test_grafana_listen(self):
+    self.assertEqual(
+        [
+            c.laddr for c in self.process_dict['grafana'].connections()
+            if c.status == 'LISTEN'
+        ],
+        [(self._ipv6_address, 8180)],
+    )
+
+  def test_influxdb_listen(self):
+    self.assertEqual(
+        sorted([
+            c.laddr for c in self.process_dict['influxdb'].connections()
+            if c.status == 'LISTEN'
+        ]),
+        [
+            (self._ipv4_address, 8088),
+            (self._ipv6_address, 8086),
+        ],
+    )
+
+  def test_telegraph_listen(self):
+    self.assertEqual(
+        [
+            c.laddr for c in self.process_dict['telegraf'].connections()
+            if c.status == 'LISTEN'
+        ],
+        [],
+    )
+
+  def test_loki_listen(self):
+    self.assertEqual(
+        sorted([
+            c.laddr for c in self.process_dict['loki'].connections()
+            if c.status == 'LISTEN'
+        ]),
+        [
+            (self._ipv4_address, 3100),
+            (self._ipv4_address, 9095),
+        ],
+    )
+
+  def test_promtail_listen(self):
+    self.assertEqual(
+        sorted([
+            c.laddr for c in self.process_dict['promtail'].connections()
+            if c.status == 'LISTEN'
+        ]),
+        [
+            (self._ipv4_address, 19080),
+            (self._ipv4_address, 19095),
+        ],
+    )