1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
|
#######
Logging
#######
************
Installation
************
Currently, we use the `sample configuration`_ in Istio to install fluentd::
cd clover/logging
First, install logging stack Elasticsearch, Fluentd and Kibana::
kubectl apply -f install/logging-stack.yaml
Note that, it must be done in separated steps. If you run ``kubectl apply -f
install`` instead, the mixer adapter may fail to intialize because the target
service can not be found. You may find an error message from mixer container::
2018-05-09T02:43:14.435156Z error Unable to initialize adapter:
snapshot='6', handler='handler.fluentd.istio-system', adapter='fluentd',
err='adapter instantiation error: dial tcp: lookup fluentd-es.logging on
10.96.0.10:53: no such host'.
Then configure fluentd for istio::
kubectl apply -f install/fluentd-istio.yaml
Configure fluentd for node level logging::
kubectl apply -f install/fluentd-daemonset-elasticsearch-rbac.yaml
.. _sample configuration: https://istio.io/docs/tasks/telemetry/fluentd.html
********
Validate
********
The scripts in ``clover/logging`` validates fluentd installation::
python clover/logging/validate.py
It validates the installation with the following criterias
#. existence of fluented pod
#. fluentd input is configured correctly
#. TBD
**************************
Understanding how it works
**************************
In clover stack, Istio is configured to automatically gather logs for services
in a mesh. More specificly, it is configured in `Mixer`_::
- when to log
- what to log
- where to log
.. _Mixer: https://istio.io/docs/concepts/policy-and-control/mixer.html
When to log
===========
Istio defines when to log by creating a custom resource ``rule``. For example:
.. code-block:: yaml
apiVersion: "config.istio.io/v1alpha2"
kind: rule
metadata:
name: newlogtofluentd
namespace: istio-system
spec:
match: "true" # match for all requests
actions:
- handler: handler.fluentd
instances:
- newlog.logentry
This rule specifies that all instances of ``newlog.logentry`` that matches the
expression will be handled by the specified handler ``handler.fluentd``. We
shall explain ``instances`` and ``handler`` later. The expression ``true`` means
whenever a request arrive at Mixer, it will trigger the actions defined belows.
``rule`` is a custom resource definition from `Istio installation`_.
.. code-block:: yaml
# Rule to send logentry instances to the fluentd handler
kind: CustomResourceDefinition
apiVersion: apiextensions.k8s.io/v1beta1
metadata:
name: rules.config.istio.io
labels:
package: istio.io.mixer
istio: core
spec:
group: config.istio.io
names:
kind: rule
plural: rules
singular: rule
scope: Namespaced
version: v1alpha2
.. _Istio installation: https://github.com/istio/istio/blob/master/install/kubernetes/templates/istio-mixer.yaml.tmpl
What to log
===========
The instance defines what content to be logged.
> A (request) instance is the result of applying request attributes to the
> template mapping. The mapping is specified as an instance configuration.
For example:
.. code-block:: yaml
# Configuration for logentry instances
apiVersion: "config.istio.io/v1alpha2"
kind: logentry
metadata:
name: newlog
namespace: istio-system
spec:
severity: '"info"'
timestamp: request.time
variables:
source: source.labels["app"] | source.service | "unknown"
user: source.user | "unknown"
destination: destination.labels["app"] | destination.service | "unknown"
responseCode: response.code | 0
responseSize: response.size | 0
latency: response.duration | "0ms"
monitored_resource_type: '"UNSPECIFIED"'
The keys under ``spec`` should conform to the template. To learn what fields
are available and valid type, you may need to reference the corresponding
template, in this case, `Log Entry template`_.
The values of each field could be either `Istio attributes`_ or an expression.
> A given Istio deployment has a fixed vocabulary of attributes that it
> understands. The specific vocabulary is determined by the set of attribute
> producers being used in the deployment. The primary attribute producer in
> Istio is Envoy, although Mixer and services can also introduce attributes.
Refer to the `Attribute Vocabulary`_ to learn the full set.
By the way, ``logentry`` is also a custom resource definition created by Istio.
.. _Istio attributes: https://istio.io/docs/concepts/policy-and-control/attributes.html
.. _Attribute Vocabulary: https://istio.io/docs/reference/config/mixer/attribute-vocabulary.html
.. _Log Entry template: https://istio.io/docs/reference/config/template/logentry.html
Where to log
============
For log, the handler defines where these information will be handled, in this
example, a fluentd daemon on fluentd-es.logging:24224.
.. code-block:: yaml
# Configuration for a fluentd handler
apiVersion: "config.istio.io/v1alpha2"
kind: fluentd
metadata:
name: handler
namespace: istio-system
spec:
address: "fluentd-es.logging:24224"
In this example, handlers (``handler.fluentd``) configure `Adapters`_
(``fluentd``) to handle the data delivered from the created instances
(``newlog.logentry``).
An adapter only accepts instance of specified kind. For example,
`fluentd adapter`_ accepts logentry but not other kinds.
.. _Adapters: https://istio.io/docs/concepts/policy-and-control/mixer.html#adapters
.. _fluentd adapter: https://istio.io/docs/reference/config/adapters/fluentd.html
|
