From 92b526db8131eae05a5787d3ebe9f5ee59fe5366 Mon Sep 17 00:00:00 2001
From: Pere Diaz Bou <pdiazbou@redhat.com>
Date: Tue, 26 Oct 2021 14:42:17 +0200
Subject: [PATCH] monitoring/grafana: Grafana query tester

Signed-off-by: Pere Diaz Bou <pdiazbou@redhat.com>
(cherry picked from commit 44d3e4c264506154373ffaeb13d6c924c580e6b5)
---
 monitoring/grafana/dashboards/.pylintrc       |   1 +
 monitoring/grafana/dashboards/CMakeLists.txt  |   6 +-
 .../grafana/dashboards/requirements-lint.txt  |  18 ++
 .../grafana/dashboards/tests/__init__.py      | 187 ++++++++++++++++++
 .../dashboards/tests/features/__init__.py     |   0
 .../tests/features/ceph-cluster.feature       |  10 +
 .../dashboards/tests/features/environment.py  | 135 +++++++++++++
 .../tests/features/hosts_overview.feature     |  28 +++
 .../dashboards/tests/features/self.feature    |  68 +++++++
 .../tests/features/steps/__init__.py          |   1 +
 .../grafana/dashboards/tests/requirements.txt |  12 ++
 monitoring/grafana/dashboards/tests/util.py   |  68 +++++++
 monitoring/grafana/dashboards/tox.ini         |  24 ++-
 .../dashboard/ci/check_grafana_dashboards.py  |  16 ++
 14 files changed, 571 insertions(+), 3 deletions(-)
 create mode 120000 monitoring/grafana/dashboards/.pylintrc
 create mode 100644 monitoring/grafana/dashboards/requirements-lint.txt
 create mode 100644 monitoring/grafana/dashboards/tests/__init__.py
 create mode 100644 monitoring/grafana/dashboards/tests/features/__init__.py
 create mode 100644 monitoring/grafana/dashboards/tests/features/ceph-cluster.feature
 create mode 100644 monitoring/grafana/dashboards/tests/features/environment.py
 create mode 100644 monitoring/grafana/dashboards/tests/features/hosts_overview.feature
 create mode 100644 monitoring/grafana/dashboards/tests/features/self.feature
 create mode 100644 monitoring/grafana/dashboards/tests/features/steps/__init__.py
 create mode 100644 monitoring/grafana/dashboards/tests/requirements.txt
 create mode 100644 monitoring/grafana/dashboards/tests/util.py

diff --git a/monitoring/grafana/dashboards/.pylintrc b/monitoring/grafana/dashboards/.pylintrc
new file mode 120000
index 0000000000000..aa04b020cb4c1
--- /dev/null
+++ b/monitoring/grafana/dashboards/.pylintrc
@@ -0,0 +1 @@
+../../../src/pybind/mgr/dashboard/.pylintrc
\ No newline at end of file
diff --git a/monitoring/grafana/dashboards/CMakeLists.txt b/monitoring/grafana/dashboards/CMakeLists.txt
index 65037cbd843ad..091f6ceeb2834 100644
--- a/monitoring/grafana/dashboards/CMakeLists.txt
+++ b/monitoring/grafana/dashboards/CMakeLists.txt
@@ -12,7 +12,9 @@ endif()
 
 if(WITH_GRAFANA)
   include(AddCephTest)
-  add_tox_test(grafana TOX_ENVS grafonnet-check)
+  add_tox_test(grafana-check TOX_ENVS grafonnet-check)
+  add_tox_test(grafana-query-test TOX_ENVS promql-query-test)
+  add_tox_test(grafana-lint TOX_ENVS lint)
   set(ver 0.1.0)
   set(name grafonnet-lib)
   include(ExternalProject)
@@ -27,7 +29,7 @@ if(WITH_GRAFANA)
     ${name})
   ExternalProject_Get_Property(${name} SOURCE_DIR)
   set_property(
-    TEST run-tox-grafana
+    TEST run-tox-grafana-check run-tox-grafana-query-test run-tox-grafana-lint
     APPEND
     PROPERTY ENVIRONMENT
     GRAFONNET_PATH=${SOURCE_DIR}/grafonnet)
diff --git a/monitoring/grafana/dashboards/requirements-lint.txt b/monitoring/grafana/dashboards/requirements-lint.txt
new file mode 100644
index 0000000000000..f9a3c772f79dd
--- /dev/null
+++ b/monitoring/grafana/dashboards/requirements-lint.txt
@@ -0,0 +1,18 @@
+attrs==21.2.0
+behave==1.2.6
+py==1.10.0
+pyparsing==2.4.7
+PyYAML==6.0
+types-PyYAML==6.0.0
+typing-extensions==3.10.0.2
+termcolor==1.1.0
+types-termcolor==1.1.2
+dataclasses==0.6
+types-dataclasses==0.6.1
+six==1.16.0
+toml==0.10.2
+pylint==2.6.0
+isort==5.10.0
+mypy==0.910
+mypy-extensions==0.4.3
+prettytable==2.4.0
diff --git a/monitoring/grafana/dashboards/tests/__init__.py b/monitoring/grafana/dashboards/tests/__init__.py
new file mode 100644
index 0000000000000..204a5f0d721b2
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/__init__.py
@@ -0,0 +1,187 @@
+import re
+import subprocess
+import sys
+import tempfile
+from dataclasses import asdict, dataclass, field
+from typing import Any, List
+
+import yaml
+
+
+@dataclass
+class InputSeries:
+    series: str = ''
+    values: str = ''
+
+@dataclass
+class ExprSample:
+    labels: str = ''
+    value: float = -1
+
+@dataclass
+class PromqlExprTest:
+    expr: str = ''
+    eval_time: str = '1m'
+    exp_samples: List[ExprSample] = field(default_factory=list)
+
+@dataclass
+class Test:
+    interval: str = '1m'
+    input_series: List[InputSeries] = field(default_factory=list)
+    promql_expr_test: List[PromqlExprTest] = field(default_factory=list)
+
+
+@dataclass
+class TestFile:
+    evaluation_interval: str = '1m'
+    tests: List[Test] = field(default_factory=list)
+
+
+class PromqlTest:
+    """
+    Base class to provide prometheus query test capabilities. After setting up
+    the query test with its input and expected output it's expected to run promtool.
+
+    https://prometheus.io/docs/prometheus/latest/configuration/unit_testing_rules/#test-yml
+
+    The workflow of testing would be something like:
+
+        # add prometheus query to test
+        self.set_expression('bonding_slaves > 0')
+
+        # add some prometheus input series
+        self.add_series('bonding_slaves{master="bond0"}', '2')
+        self.add_series('bonding_slaves{master="bond1"}', '3')
+        self.add_series('node_network_receive_bytes{instance="127.0.0.1",
+            device="eth1"}', "10 100 230 22")
+
+        # expected output of the query
+        self.add_exp_samples('bonding_slaves{master="bond0"}', 2)
+        self.add_exp_samples('bonding_slaves{master="bond1"}', 3)
+
+        # at last, always call promtool with:
+        self.assertTrue(self.run_promtool())
+        # assertTrue means it expect promtool to succeed
+    """
+
+    def __init__(self):
+        self.test_output_file = tempfile.NamedTemporaryFile('w+')
+
+        self.test_file = TestFile()
+        self.test = Test()
+        self.promql_expr_test = PromqlExprTest()
+        self.test.promql_expr_test.append(self.promql_expr_test)
+        self.test_file.tests.append(self.test)
+
+        self.variables = {}
+
+    def __del__(self):
+        self.test_output_file.close()
+
+
+    def set_evaluation_interval(self, interval: int, unit: str = 'm') -> None:
+        """
+        Set the evaluation interval of the time series
+
+        Args:
+            interval (int): number of units.
+            unit (str): unit type: 'ms', 's', 'm', etc...
+        """
+        self.test_file.evaluation_interval = f'{interval}{unit}'
+
+    def set_interval(self, interval: int, unit: str = 'm') -> None:
+        """
+        Set the duration of the time series
+
+        Args:
+            interval (int): number of units.
+            unit (str): unit type: 'ms', 's', 'm', etc...
+        """
+        self.test.interval = f'{interval}{unit}'
+
+    def set_expression(self, expr: str) -> None:
+        """
+        Set the prometheus expression/query used to filter data.
+
+        Args:
+             expr(str): expression/query.
+        """
+        self.promql_expr_test.expr = expr
+
+    def add_series(self, series: str, values: str) -> None:
+        """
+        Add a series to the input.
+
+        Args:
+             series(str): Prometheus series.
+             Notation: '<metric name>{<label name>=<label value>, ...}'
+             values(str): Value of the series.
+        """
+        input_series = InputSeries(series=series, values=values)
+        self.test.input_series.append(input_series)
+
+    def set_eval_time(self, eval_time: int, unit: str = 'm') -> None:
+        """
+        Set the time when the expression will be evaluated
+
+        Args:
+             interval (int): number of units.
+             unit (str): unit type: 'ms', 's', 'm', etc...
+        """
+        self.promql_expr_test.eval_time = f'{eval_time}{unit}'
+
+    def add_exp_samples(self, sample: str, values: Any) -> None:
+        """
+        Add an expected sample/output of the query given the series/input
+
+        Args:
+             sample(str): Expected sample.
+             Notation: '<metric name>{<label name>=<label value>, ...}'
+             values(Any): Value of the sample.
+        """
+        expr_sample = ExprSample(labels=sample, value=values)
+        self.promql_expr_test.exp_samples.append(expr_sample)
+
+    def set_variable(self, variable: str, value: str):
+        """
+        If a query makes use of grafonnet variables, for example
+        '$osd_hosts', you should change this to a real value. Example:
+
+
+        > self.set_expression('bonding_slaves{master="$osd_hosts"} > 0')
+        > self.set_variable('osd_hosts', '127.0.0.1')
+        > print(self.query)
+        > bonding_slaves{master="127.0.0.1"} > 0
+
+        Args:
+             variable(str): Variable name
+             value(str): Value to replace variable with
+
+        """
+        self.variables[variable] = value
+
+    def run_promtool(self):
+        """
+        Run promtool to test the query after setting up the input, output
+        and extra parameters.
+
+        Returns:
+             bool: True if successful, False otherwise.
+        """
+
+        for variable, value in self.variables.items():
+            expr = self.promql_expr_test.expr
+            new_expr = re.sub(r'\${0}'.format(variable), str(value), expr)
+            self.set_expression(new_expr)
+
+        test_as_dict = asdict(self.test_file)
+        yaml.dump(test_as_dict, self.test_output_file)
+
+        args = f'promtool test rules {self.test_output_file.name}'.split()
+        try:
+            subprocess.run(args, check=True)
+            return True
+        except subprocess.CalledProcessError as process_error:
+            print(yaml.dump(test_as_dict))
+            print(process_error.stderr)
+            return False
diff --git a/monitoring/grafana/dashboards/tests/features/__init__.py b/monitoring/grafana/dashboards/tests/features/__init__.py
new file mode 100644
index 0000000000000..e69de29bb2d1d
diff --git a/monitoring/grafana/dashboards/tests/features/ceph-cluster.feature b/monitoring/grafana/dashboards/tests/features/ceph-cluster.feature
new file mode 100644
index 0000000000000..f8b580566f451
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/features/ceph-cluster.feature
@@ -0,0 +1,10 @@
+Feature: Ceph Cluster Dashboard
+
+Scenario: "Test total PG States"
+  Given the following series:
+    | metrics | values |
+    | ceph_pg_total{foo="var"} | 10 100 |
+    | ceph_pg_total{foo="bar"} | 20 200 |
+  Then Grafana panel `PG States` with legend `Total` shows:
+    | metrics | values |
+    | {} | 300 |
\ No newline at end of file
diff --git a/monitoring/grafana/dashboards/tests/features/environment.py b/monitoring/grafana/dashboards/tests/features/environment.py
new file mode 100644
index 0000000000000..8509b9d97e8e2
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/features/environment.py
@@ -0,0 +1,135 @@
+# type: ignore[no-redef]
+# pylint: disable=E0611,W0613,E0102
+import copy
+
+from behave import given, then, when
+from prettytable import PrettyTable
+
+from tests import PromqlTest
+from tests.util import get_dashboards_data, resolve_time_and_unit
+
+
+class GlobalContext:
+    def __init__(self):
+        self.tested_queries_count = 0
+        self.promql_expr_test = None
+        self.data = get_dashboards_data()
+        self.query_map = self.data['queries']
+
+    def reset_promql_test(self):
+        self.promql_expr_test = PromqlTest()
+        self.promql_expr_test.variables = copy.copy(self.data['variables'])
+
+    def print_query_stats(self):
+        total = len(self.query_map)
+        table = PrettyTable()
+        table.field_names = ['Name', 'Queries', 'Tested', 'Cover']
+
+        def percent(tested, total):
+            return str(round((tested / total) * 100, 2)) + '%'
+
+        def file_name(path):
+            return path.split('/')[-1]
+
+        total = 0
+        tested = 0
+        for path, stat in self.data['stats'].items():
+            assert stat['total']
+            table.add_row([file_name(path), stat['total'], stat['tested'],
+                                     percent(stat['tested'], stat['total'])])
+            total += stat['total']
+            tested += stat['tested']
+
+        assert total
+        table.add_row(['Total', total, tested, percent(tested, total)])
+        print(table)
+
+
+global_context = GlobalContext()
+
+# Behave function overloading
+# ===========================
+
+
+def before_scenario(context, scenario):
+    global_context.reset_promql_test()
+
+
+def after_scenario(context, scenario):
+    assert global_context.promql_expr_test.run_promtool()
+
+
+def after_all(context):
+    global_context.print_query_stats()
+
+
+@given("the following series")
+def step_impl(context):
+    for row in context.table:
+        metric = row['metrics']
+        value = row['values']
+        global_context.promql_expr_test.add_series(metric, value)
+
+
+@when('evaluation interval is `{interval}`')
+def step_impl(context, interval):
+    interval_without_unit, unit = resolve_time_and_unit(interval)
+    if interval_without_unit is None:
+        raise ValueError(f'Invalid interval time: {interval_without_unit}. ' +
+                           'A valid time looks like "1m" where you have a number plus a unit')
+    global_context.promql_expr_test.set_evaluation_interval(interval_without_unit, unit)
+
+
+@when('interval is `{interval}`')
+def step_impl(context, interval):
+    interval_without_unit, unit = resolve_time_and_unit(interval)
+    if interval_without_unit is None:
+        raise ValueError(f'Invalid interval time: {interval_without_unit}. ' +
+                           'A valid time looks like "1m" where you have a number plus a unit')
+    global_context.promql_expr_test.set_interval(interval_without_unit, unit)
+
+
+@when('evaluation time is `{eval_time}`')
+def step_impl(context, eval_time):
+    eval_time_without_unit, unit = resolve_time_and_unit(eval_time)
+    if eval_time_without_unit is None:
+        raise ValueError(f'Invalid evalution time: {eval_time}. ' +
+                           'A valid time looks like "1m" where you have a number plus a unit')
+    global_context.promql_expr_test.set_eval_time(eval_time_without_unit, unit)
+
+
+@when('variable `{variable}` is `{value}`')
+def step_impl(context, variable, value):
+    global_context.promql_expr_test.set_variable(variable, value)
+
+
+@then('Grafana panel `{panel_name}` with legend `{legend}` shows')
+def step_impl(context, panel_name, legend):
+    """
+    This step can have an empty legend. As 'behave' doesn't provide a way
+    to say it's empty we use EMPTY to mark as empty.
+    """
+    if legend == "EMPTY":
+        legend = ''
+    query_id = panel_name + '-' + legend
+    if query_id not in global_context.query_map:
+        raise KeyError((f'Query with legend {legend} in panel "{panel_name}"'
+                           'couldn\'t be found'))
+
+    expr = global_context.query_map[query_id]['query']
+    global_context.promql_expr_test.set_expression(expr)
+    for row in context.table:
+        metric = row['metrics']
+        value = row['values']
+        global_context.promql_expr_test.add_exp_samples(metric, float(value))
+    path = global_context.query_map[query_id]['path']
+    global_context.data['stats'][path]['tested'] += 1
+
+
+@then('query `{query}` produces')
+def step_impl(context, query):
+    global_context.promql_expr_test.set_expression(query)
+    for row in context.table:
+        metric = row['metrics']
+        value = row['values']
+        global_context.promql_expr_test.add_exp_samples(metric, float(value))
diff --git a/monitoring/grafana/dashboards/tests/features/hosts_overview.feature b/monitoring/grafana/dashboards/tests/features/hosts_overview.feature
new file mode 100644
index 0000000000000..e3ddf57bb2bf2
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/features/hosts_overview.feature
@@ -0,0 +1,28 @@
+Feature: Hosts Overview Dashboard
+
+Scenario: "Test network load succeeds"
+  Given the following series:
+    | metrics | values |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+  When variable `osd_hosts` is `127.0.0.1`
+  Then Grafana panel `Network Load` with legend `EMPTY` shows:
+    | metrics | values |
+    | {} | 6 |
+
+Scenario: "Test network load with bonding succeeds"
+  Given the following series:
+    | metrics | values |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth1"} | 10 100 200 |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth2"} | 10 100 200 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 100 200 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 100 200 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="bond0"} | 20 200 300 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="bond0"} | 20 200 300 |
+    | bonding_slaves{instance="127.0.0.1", master="bond0"} | 2 |
+  When variable `osd_hosts` is `127.0.0.1`
+  Then Grafana panel `Network Load` with legend `EMPTY` shows:
+    | metrics | values |
+    | {} | 6 |
diff --git a/monitoring/grafana/dashboards/tests/features/self.feature b/monitoring/grafana/dashboards/tests/features/self.feature
new file mode 100644
index 0000000000000..2b44ce0dcecb1
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/features/self.feature
@@ -0,0 +1,68 @@
+Feature: Test tester
+
+Scenario: "Simple query works"
+  Given the following series:
+    | metrics | values |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="192.168.100.2", device="bond0"} | 20 200 |
+    | node_network_transmit_bytes{instance="192.168.100.1", device="bond0"} | 20 200 |
+    | bonding_slaves{instance="127.0.0.1", master="bond0"} | 2 |
+  Then query `node_network_transmit_bytes{instance="127.0.0.1"} > 0` produces:
+    | metrics | values |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 100 |
+
+Scenario: "Query with evaluation time"
+  Given the following series:
+    | metrics | values |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="192.168.100.2", device="bond0"} | 20 200 |
+    | node_network_transmit_bytes{instance="192.168.100.1", device="bond0"} | 20 200 |
+    | bonding_slaves{instance="127.0.0.1", master="bond0"} | 2 |
+  When evaluation time is `0m`
+  Then query `node_network_transmit_bytes{instance="127.0.0.1"} > 0` produces:
+    | metrics | values |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 |
+
+Scenario: "Query with evaluation time and variable value"
+  Given the following series:
+    | metrics | values |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 100 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 100 |
+    | node_network_transmit_bytes{instance="192.168.100.2", device="bond0"} | 20 200 |
+    | node_network_transmit_bytes{instance="192.168.100.1", device="bond0"} | 20 200 |
+    | bonding_slaves{instance="127.0.0.1", master="bond0"} | 2 |
+  When evaluation time is `0m`
+  And variable `osd_hosts` is `127.0.0.1`
+  Then query `node_network_transmit_bytes{instance="$osd_hosts"} > 0` produces:
+    | metrics | values |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 |
+
+Scenario: "Query with interval time"
+  Given the following series:
+    | metrics | values |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth1"} | 10 100 200 |
+    | node_network_receive_bytes{instance="127.0.0.1", device="eth2"} | 10 100 200 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 10 100 200 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 10 100 200 |
+    | node_network_transmit_bytes{instance="192.168.100.2", device="bond0"} | 20 200 300 |
+    | node_network_transmit_bytes{instance="192.168.100.1", device="bond0"} | 20 200 300 |
+    | bonding_slaves{instance="127.0.0.1", master="bond0"} | 2 |
+  When evaluation time is `2h`
+  And evaluation interval is `1h`
+  And interval is `1h`
+  And variable `osd_hosts` is `127.0.0.1`
+  Then query `node_network_transmit_bytes{instance="$osd_hosts"} > 0` produces:
+    | metrics | values |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth1"} | 200 |
+    | node_network_transmit_bytes{instance="127.0.0.1", device="eth2"} | 200 |
\ No newline at end of file
diff --git a/monitoring/grafana/dashboards/tests/features/steps/__init__.py b/monitoring/grafana/dashboards/tests/features/steps/__init__.py
new file mode 100644
index 0000000000000..0b90f46f2c38f
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/features/steps/__init__.py
@@ -0,0 +1 @@
+# This file and steps files is needed even if its empty because of 'behave' :(
diff --git a/monitoring/grafana/dashboards/tests/requirements.txt b/monitoring/grafana/dashboards/tests/requirements.txt
new file mode 100644
index 0000000000000..8ad130e5bbe61
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/requirements.txt
@@ -0,0 +1,12 @@
+attrs==21.2.0
+behave==1.2.6
+py==1.10.0
+pyparsing==2.4.7
+PyYAML==6.0
+types-PyYAML==6.0.0
+typing-extensions==3.10.0.2
+termcolor==1.1.0
+types-termcolor==1.1.2
+dataclasses==0.6
+types-dataclasses==0.6.1
+prettytable==2.4.0
diff --git a/monitoring/grafana/dashboards/tests/util.py b/monitoring/grafana/dashboards/tests/util.py
new file mode 100644
index 0000000000000..4f09e9edd3b0f
--- /dev/null
+++ b/monitoring/grafana/dashboards/tests/util.py
@@ -0,0 +1,68 @@
+import json
+from pathlib import Path
+from typing import Any, Dict, Tuple, Union
+
+from termcolor import cprint
+
+UNITS = ['ms', 's', 'm', 'h', 'd', 'w', 'y']
+
+
+def resolve_time_and_unit(time: str) -> Union[Tuple[int, str], Tuple[None, None]]:
+    """
+    Divide time with its unit and return a tuple like (10, 'm')
+    Return None if its and invalid prometheus time
+    Valid units are inside UNITS.
+    """
+    if time[-1] in UNITS:
+        return int(time[:-1]), time[-1]
+    if time[-2:] in UNITS:
+        return int(time[:-2]), time[-2:]
+    return None, None
+
+
+def get_dashboards_data() -> Dict[str, Any]:
+    data: Dict[str, Any] = {'queries': {}, 'variables': {}, 'stats': {}}
+    for file in Path(__file__).parent.parent.glob('*.json'):
+        with open(file, 'r') as f:
+            dashboard_data = json.load(f)
+            data['stats'][str(file)] = {'total': 0, 'tested': 0}
+            add_dashboard_queries(data, dashboard_data, str(file))
+            add_dashboard_variables(data, dashboard_data)
+    return data
+
+
+def add_dashboard_queries(data: Dict[str, Any], dashboard_data: Dict[str, Any], path: str) -> None:
+    """
+    Grafana panels can have more than one target/query, in order to identify each
+    query in the panel we append the "legendFormat" of the target to the panel name.
+    format: panel_name-legendFormat
+    """
+    if 'panels' not in dashboard_data:
+        return
+    for panel in dashboard_data['panels']:
+        if (
+                'title' in panel
+                and 'targets' in panel
+                and len(panel['targets']) > 0
+                and 'expr' in panel['targets'][0]
+        ):
+            for target in panel['targets']:
+                title = panel['title']
+                legend_format = target['legendFormat'] if 'legendFormat' in target else ""
+                query_id = title + '-' + legend_format
+                if query_id in data['queries']:
+                    # NOTE: If two or more panels have the same name and legend it
+                    # might suggest a refactoring is needed or add something else
+                    # to identify each query.
+                    cprint((f'WARNING: Query in panel "{title}" with legend "{legend_format}"'
+                                       ' already exists'), 'yellow')
+                data['queries'][query_id] = {'query': target['expr'], 'path': path}
+                data['stats'][path]['total'] += 1
+
+
+def add_dashboard_variables(data: Dict[str, Any], dashboard_data: Dict[str, Any]) -> None:
+    if 'templating' not in dashboard_data or 'list' not in dashboard_data['templating']:
+        return
+    for variable in dashboard_data['templating']['list']:
+        if 'name' in variable:
+            data['variables'][variable['name']] = 'UNSET VARIABLE'
diff --git a/monitoring/grafana/dashboards/tox.ini b/monitoring/grafana/dashboards/tox.ini
index 10aeb9f38aab2..382952c5b1bee 100644
--- a/monitoring/grafana/dashboards/tox.ini
+++ b/monitoring/grafana/dashboards/tox.ini
@@ -1,5 +1,5 @@
 [tox]
-envlist = grafonnet-{check,fix}
+envlist = grafonnet-{check,fix},lint,promql-query-test
 skipsdist = true
 
 [grafonnet]
@@ -20,3 +20,25 @@ passenv = GRAFONNET_PATH
 commands =
      check: bash test-jsonnet.sh
      fix: jsonnet -m . jsonnet/grafana_dashboards.jsonnet
+
+
+[testenv:lint]
+description =
+     Run linters
+deps =
+     -rrequirements-lint.txt
+setenv =
+commands =
+    pylint --rcfile=.pylintrc tests
+    mypy tests
+    isort tests
+
+[testenv:promql-query-test]
+description =
+     Run promtool unit testing on grafana queries.
+deps =
+     -rtests/requirements.txt
+depends = grafonnet-check
+setenv =
+commands =
+    behave tests/features 
diff --git a/src/pybind/mgr/dashboard/ci/check_grafana_dashboards.py b/src/pybind/mgr/dashboard/ci/check_grafana_dashboards.py
index 47e553c84e585..03b31a45e173e 100644
--- a/src/pybind/mgr/dashboard/ci/check_grafana_dashboards.py
+++ b/src/pybind/mgr/dashboard/ci/check_grafana_dashboards.py
@@ -76,6 +76,22 @@ def get_grafana_dashboards(base_dir):
             with open(json_file) as f:
                 dashboard_config = json.load(f)
                 uid = dashboard_config.get('uid')
+                # if it's not a grafana dashboard, skip checks
+                # Fields in a dasbhoard:
+                # https://grafana.com/docs/grafana/latest/dashboards/json-model/#json-fields
+                expected_fields = [
+                    'id', 'uid', 'title', 'tags', 'style', 'timezone', 'editable',
+                    'hideControls', 'graphTooltip', 'panels', 'time', 'timepicker',
+                    'templating', 'annotations', 'refresh', 'schemaVersion', 'version', 'links',
+                ]
+                not_a_dashboard = False
+                for field in expected_fields:
+                    if field not in dashboard_config:
+                        not_a_dashboard = True
+                        break
+                if not_a_dashboard:
+                    continue
+
                 assert dashboard_config['id'] is None, \
                     "'id' not null: '{}'".format(dashboard_config['id'])
 
-- 
2.39.5