Skip to content
master
Go to file
Code

Latest commit

…ranch (#3593)

This adds:
- API support for mesh discovery and home cluster identification
  - "Home cluster" means: the cluster where the "adjacent" control plane is running. "Adjacent control plane" means the 
     control plane that Kiali is observing/managing.
  - Remote clusters are discovered by fetching the list of "remote secrets" that are present in the istio namespace.
    Apparently, remote secrets are labeled with istio/multiCluster=true (at least, this is the label that presets the `istioctl x 
    create-remote-secret` command). So, this is fetching secrets with that label and assuming they contain the kubeconfig 
    files required by istio to contact the remote API Server and enable discovery.
- Some Config additions for mesh/multi-cluster config
- Graph support for clusters, cluster and namespace boxing
  - add support for "cluster" queryParam for node graph requests. this mantains the REST API while also allowing flexibility 
    to graph a node cross-cluster or in a cluster-specific way.
  - add Cluster to NodeData and DestServices
  - add Cluster to the JSON sorting
  - change "Group" to "Box" to consolidate terminology and avoid the overused 'group' terminology.
  - add some handling for bad istio telemetry (see istio/istio#29373)
  - comment out old MC telem massaging as it should be outdated

#3592

Co-authored-by: Edgar Hernández <ehernand@redhat.com>
6ec37a5

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
Feb 3, 2018

README.adoc

Kiali Tweet about Kiali

Build Status Apache 2.0 license

Introduction

Kiali provides answers to the questions: What microservices are part of my Istio service mesh and how are they connected?

Kiali Graph

Table of contents

Description

A Microservice Architecture breaks up the monolith into many smaller pieces that are composed together. Patterns to secure the communication between services like fault tolerance (via timeout, retry, circuit breaking, etc.) have come up as well as distributed tracing to be able to see where calls are going.

A service mesh can now provide these services on a platform level and frees the application writers from those tasks. Routing decisions are done at the mesh level.

Kiali works with Istio to visualise the service mesh topology, features like circuit breakers or request rates.

Kiali also includes an integration with Jaeger Tracing to provide distributed tracing out of the box.

Getting Started

This README is targeted mainly toward Kiali developers. If you are not a developer but want to learn more about Kiali, you will find the Kiali documentation more helpful. For some instructions on running Kiali without having to git clone this repository or build anything, please see the Getting Started page.

Testing on a OpenShift 4.x Cluster

To install a OpenShift 4.x cluster in AWS, use this hack script.

Run the hack script with --help for usage details. Here is a brief synopsis:

  • aws-openshift.sh -kuca true create : Create an AWS cluster that includes Service Mesh, Kiali, Jaeger, ElasticSearch

  • aws-openshift.sh status : Provides details about your AWS cluster

  • aws-openshift.sh oc-env : Provides details about how to set up the oc client in your shell so it can access your AWS cluster

  • aws-openshift.sh destroy : Destroy the AWS cluster

  • aws-openshift.sh bi-install : Install BookInfo demo in the AWS cluster

  • aws-openshift.sh sm-uninstall : Uninstall the Service Mesh and all its components that were previously installed via create or sm-install

  • aws-openshift.sh sm-install : Re-install the Service Mesh and all its components

  • aws-openshift.sh k-uninstall : Uninstall the Kiali component that was installed via create or sm-install

Here’s a way you can work with AWS to do Kiali development:

  • aws-openshift.sh -kuca true create to create your cluster with Service Mesh. -kuca true tells the script to create a cluster admin user "kiali" with password "kiali". This takes roughly 30 minutes to complete.

  • aws-openshift.sh status to show you details about your cluster, including "oc login" details as well as the console URL among other things

  • If you want to build and deploy your own Kiali:

    • First must uninstall the one that was installed when you created the cluster. To do this, run aws-openshift.sh k-uninstall

    • Next build Kiali and the Kiali operator normally (see below for details on that)

    • Push the dev builds of your Kiali and Kiali operator into the AWS cluster’s image registry via make cluster-push.

    • Create the new Kiali Operator in your AWS cluster via make operator-create

    • Create the new Kiali in your AWS cluster via make kiali-create

    • At this point, your dev build of Kiali should be starting up. Wait for it to come up and then you can access the Kiali UI normally.

Container Image

The container images for Kiali and the Kiali Operator are published on Quay.io in the kiali organization.

See the LICENSE file.

Building

Note
These build instructions assume you have the following installed on your system: (1) Go Programming Language, (2) git, (3) Docker or Podman , (4) NPM, and (5) make. If you are using podman instead of docker, pass the environment variable DORP=podman when executing make. To run Kiali in a cluster after you build it, it is assumed you have a running OpenShift or Minikube environment available to you.
Note
Currently, Kiali releases are built using Go 1.14. Although Kiali may build correctly using other versions of Go, it’s suggested to use version 1.14 for development to ensure replicatable builds. Makefiles will require this minimum version of Go.

To build Kiali:

  • Clone this repository inside a GOPATH. These instructions will use the example GOPATH of "/source/kiali/kiali" but you can use whatever you want. Just change the first line of the below instructions to use your GOPATH.

export GOPATH=/source/kiali/kiali
mkdir -p $GOPATH
cd $GOPATH
mkdir -p src/github.com/kiali
cd src/github.com/kiali
git clone git@github.com:kiali/kiali
git clone git@github.com:kiali/kiali-operator kiali/operator
git clone git@github.com:kiali/helm-charts kiali/helm-charts
export PATH=${PATH}:${GOPATH}/bin
  • Build Kiali

cd ${GOPATH}/src/github.com/kiali/kiali
make build
  • At this point you can run the Kiali tests

cd ${GOPATH}/src/github.com/kiali/kiali
make test

Running

Note
If you want to quickly get up and running to play with Kiali and do not wish to git clone the repository or build anything, see https://www.kiali.io/gettingstarted for instructions.
Note
Before deploying and running Kiali, you must first install and deploy Istio. There are a few places that you can reference in order to learn how to do this. We recommend using Maistra which is a variant of Istio. If you choose to use Istio, make sure to follow the OpenShift preparation steps. Also, check the version requirements on our website to read notes about Istio and Maistra compatibility. There is a install-istio-via-istioctl.sh hack script that can install Istio for you.

Running on OpenShift or Kubernetes/Minikube

By default, the make targets used to push and deploy Kiali images will assume you have an OpenShift 4.x cluster installed and running and that oc is found in your $PATH.

If you wish to push and deploy Kiali to an installed and running Kubernetes (via Minikube) environment, pass the environment variable CLUSTER_TYPE=minikube to the make commands and make sure kubectl is found in your $PATH.

If you have neither minikube nor a remote OpenShift cluster, you can pass the environment variable CLUSTER_TYPE=local to the make commands and make sure you have either oc or kubectl in your $PATH. This requires your Kubernetes cluster to be able to pull from your local image repository.

In order to deploy on Minikube using the below instructions, and to be able to access the deployed services, you must ensure you have the Registry and Ingress addons. To do this, ensure you run minikube addons enable registry and minikube addons enable ingress and add kiali as a hostname in your /etc/hosts via something like this command: echo "$(minikube ip) kiali" | sudo tee -a /etc/hosts

If you are using docker, make sure the endpoint for your OpenShift image registry or Kubernetes image registry is configured as an insecure registry in your docker daemon config.

Building the Container Image

Create the Kiali container images through the "container-build" make target:

cd ${GOPATH}/src/github.com/kiali/kiali
make container-build

This will build the Kiali image.

Generated container image will be consistant with the host machine (either x86_64 or aarch64).

Pushing Kiali operator and Kiali images to your cluster

cd ${GOPATH}/src/github.com/kiali/kiali
make cluster-push

This actually invokes the container-build target as well, so you technically do not have to run make container-build as a separate step.

Deploying Kiali operator to your cluster

cd ${GOPATH}/src/github.com/kiali/kiali
make operator-create

Deploying the Kiali operator needs to be done only one time. After the operator is deployed, you can deploy and remove Kiali any number of times using the steps described below.

Deploying Kiali to your cluster

cd ${GOPATH}/src/github.com/kiali/kiali
make kiali-create

This creates a Kiali custom resource (CR) which instructs the Kiali operator to install Kiali. A secret will be created if needed (i.e. when the authentication strategy is set to login).

Removing Kiali from your cluster

If you want to remove Kiali, you can do so via the operator by executing the kiali-delete target. This target will also remove the secret.

cd ${GOPATH}/src/github.com/kiali/kiali
make kiali-delete

If you want to purge your cluster of all Kiali resources without going through the operator to do it, you can run the kiali-purge target instead. This does not use the operator, instead it purges all Kiali resources using the oc/kubectl command.

cd ${GOPATH}/src/github.com/kiali/kiali
make kiali-purge

Reloading Kiali image

Note
To easily facilitate development, this allows you to do make clean build cluster-push-kiali kiali-reload-image to quickly get your new Kiali build into your running cluster.

If you already have Kiali installed but you want to recreate the pod with a new container image, you can run the following command:

cd ${GOPATH}/src/github.com/kiali/kiali
make kiali-reload-image

Removing Kiali operator from your cluster

cd ${GOPATH}/src/github.com/kiali/kiali
make operator-delete
Note
After this completes, the kiali-create and kiali-delete targets will be ineffective until you run the operator-create target to re-deploy the Kiali operator again.

Running Standalone

Rarely, you may want to run Kiali outside of any cluster environment, perhaps for debugging purposes. To do this, run:

cd ${GOPATH}/src/github.com/kiali/kiali
make install
make run

The "install" target installs the Kiali executable in your GOPATH /bin directory so you can run it outside of the Makefile:

cd ${GOPATH}/src/github.com/kiali/kiali
make install
${GOPATH}/bin/kiali -config <your-config-file>

Configuration

Many configuration settings can optionally be set within the Kiali Operator custom resource (CR) file. See this example Kiali CR file that has all the configuration settings documented.

Embedding Kiali

If you want to embed Kiali in other applications, Kiali offers a simple feature called Kiosk mode. In this mode, Kiali won’t show the main header, nor the main navigation bar.

To enable Kiosk mode, you only need to add a kiosk=true URL parameter. You will need to use the full path of the page you want to embed. For example, assuming that you access Kiali through HTTPS:

  • To embed the Overview page, use https://kiali_path/overview?kiosk=true.

  • To embed the Graph page, use https://kiali_path/graph/namespaces?kiosk=true.

  • To embed the Applications list page, use https://kiali_path/applications?kiosk=true.

If the page you want to embed uses other URL arguments, you can specify any of them to preset options. For example, if you want to embed the graph of the bookinfo namespace, use the following URL: http://kiali_path/graph/namespaces?namespaces=bookinfo&kiosk=true.

Configure External Services

Jaeger

If you have Jaeger installed in a custom way that is not easily auto-detectable by Kiali, you need to change in the Kiali CR the value of the jaeger > url

apiVersion: kiali.io/v1alpha1
kind: Kiali
metadata:
  name: kiali
spec:
...
    external_services:
      jaeger:
        url: http://jaeger-query-istio-system.127.0.0.1.nip.io
...

Grafana

If you have Grafana installed in a custom way that is not easily auto-detectable by Kiali, you need to change in the Kiali CR the value of the grafana > url

apiVersion: kiali.io/v1alpha1
kind: Kiali
metadata:
  name: kiali
spec:
...
    external_services:
      grafana:
        url: http://grafana-istio-system.127.0.0.1.nip.io
...

Services API Documentation

Kiali can display API Documentation of your services. See API documentation screen capture and API type list screen capture.

Configure your services

Your services must be annotated with the type of API ('rest', 'grpc', 'graphql') and a URL to the spec of the API. If the API spec is served from the service itself, Kiali will infer the hostname and port :

apiVersion: v1
kind: Service
metadata:
  name: myservice
  annotations:
    kiali.io/api-type: rest
    kiali.io/api-spec: /v1/api-spec
spec:
...

The API spec can also be served from any http/s URL, internal or external to the cluster :

apiVersion: v1
kind: Service
metadata:
  name: petstore
  annotations:
    kiali.io/api-type: rest
    kiali.io/api-spec: https://petstore.swagger.io/v2/swagger.json
spec:
...

For now, only REST APIs have their spec displayed but we are working to support gRpc and GraphQL soon. A live console to test your APIs directly with Kiali is also being worked on.

Customize API docs annotations

You can configure Kiali to use your own annotation names with the Kiali CR

...
apidocs:
  annotations:
    api_spec_annotation_name: "my-annotation-for-api-spec"
    api_type_annotation_name: "my-annotation-for-api-type"
...

Additional Notes

Customize the UI web context root

By default Kiali UI is deployed to the top level of https://kiali-istio-system.<your_cluster_domain_or_ip>/. In some situations such as when you want to serve Kiali UI along with other apps under the same host name, e.g., example.com/kiali, example.com/app1, you can edit Kiali Config Map and provide a different value for web_root. Note: the path must begin with a / and not end with / (e.g. /kiali).

An example of custom web root:

...
server:
  web_root: /kiali
...

Running the UI Outside the Core

When developing the Kiali UI you will find it useful to run it outside of the core to make it easier to update the UI code and see the changes without having to recompile. The prefered approach for this is to use a proxy on the UI to mount the core. The process is described here.

To connect with the backend and avoid the javascript prompt requesting authentication you need to send the requests with a specific header.

X-Auth-Type-Kiali-UI: 1

The response will contain the header

WWW-Authenticate: xBasic realm="Kiali"

Otherwise the header will be

WWW-Authenticate: Basic realm="Kiali"

Running A Locally Built UI Inside the Core

If you are developing the UI on your local machine but you want to see it deployed and running inside of the core server, you can do so by setting the environment variable CONSOLE_VERSION to the value "local" when building the container image via the container-build or cluster-build target. By default, your UI’s build/ directory is assumed to be in a directory called kiali-ui that is a peer directory of the GOPATH root directory for the core server. If it is not, you can set the environment variable CONSOLE_LOCAL_DIR to the value of the path of the root directory for the UI such that $CONSOLE_LOCAL_DIR/build contains the generated build files for the UI.

For example, if your GOPATH directory for the Kiali project is /source/kiali/kiali and you have git cloned the Kiali UI repository in /source/kiali/kiali-ui then you do not need to set CONSOLE_LOCAL_DIR. You can embed your locally built console into the core container image via:

CONSOLE_VERSION=local make container-build

If you git cloned the Kiali UI repository in directory /my/git/repo and have built the UI there (such that the build files are located at /my/git/repo/build) then you can embed that locally built console into the core container image via:

CONSOLE_VERSION=local CONSOLE_LOCAL_DIR=/my/git/repo make container-build

Disabling SSL

In the provided OpenShift templates, SSL is turned on by default. If you want to turn it off, you should:

  • Remove the "tls: termination: reencrypt" option from the Kiali route

  • Remove the "identity" block, with certificate paths, from the Kiali Config Map.

  • Optionally you can also remove the annotation "service.alpha.openshift.io/serving-cert-secret-name", and the related volume that is declared and mounted in Kiali Deployment (but if you don’t, they will just be ignored).

In the provided Kubernetes templates, SSL is turned on by default. If you want to turn it off, you should:

  • Remove the "identity" block, with certificate paths, from the Kiali Config Map.

  • Optionally you can also remove the volume that is declared and mounted in Kiali Deployment (the name of the volume and mount are both "kiali-cert". If you don’t remove these, they will just be ignored.

Exposing Kiali to External Clients Using Istio Gateway

The operator will create a Route or Ingress by default (see the Kiali CR setting "deployment.ingress_enabled"). If you want to expose Kiali via Istio itself, you can create Gateway, Virtual Service, and Destination Rule resources similiar to below:

---
apiVersion: networking.istio.io/v1alpha3
kind: Gateway
metadata:
  name: kiali-gateway
  namespace: istio-system
spec:
  selector:
    istio: ingressgateway
  servers:
  - port:
      number: 80
      name: http-kiali
      protocol: HTTP
    # https://istio.io/latest/docs/reference/config/networking/gateway/#ServerTLSSettings
    tls:
      httpsRedirect: false
    hosts: [<your-host>]
  - port:
      number: 443
      name: https-kiali
      protocol: HTTPS
    tls: {}
    hosts: [<your-host>]
...
---
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: kiali-virtualservice
  namespace: istio-system
spec:
  gateways:
  - kiali-gateway
  hosts: [<your-host>]
  http:
  - route:
    - destination:
        host: kiali.istio-system.svc.cluster.local
        port:
          number: 20001
      weight: 100
...
---
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
  name: kiali-destinationrule
  namespace: istio-system
spec:
  host: kiali
  trafficPolicy:
    tls:
      mode: DISABLE
...

Experimental

Observing a Remote Cluster

There are certain use cases where Kiali needs to be deployed in one cluster (Control Plane) and observe a different cluster (Data Plane). Diagram.

Follow these steps:

1: You should have the remote central istiod with a single cluster setup running

2: Create the Kiali ClusterRole, ClusterRoleBinding, and ServiceAccount in the Data Plane cluster

3: Create a remote secret in the Control Plane, using the Data Plane ServiceAccount you just created. This allows the Control Plane to read from and modify the Data Plane

istioctl x create-remote-secret --service-account kiali-service-account --context=$DataPlane --name kiali | kubectl apply -n istio-system --context=$ControlPlane -f -

4: You will now run Kiali in the Control Plane. You need to add the remote secret to the Kiali Deployment by specifying a Volume and VolumeMount. When Kiali sees /kiali-remote-secret/kiali it will use the remote cluster’s API server instead of the local API server

spec:
  template:
    spec:
      containers:
      - volumeMounts:
        - mountPath: /kiali-remote-secret
          name: kiali-remote-secret
      volumes:
      - name: kiali-remote-secret
        secret:
          defaultMode: 420
          optional: true
          secretName: istio-remote-secret-kiali

5: Kiali now needs the Istio metrics from the sidecars. You need to run Prometheus in the Control Plane and have it scrape the metrics from an envoyMetricsService. These metrics are required:

  • istio_requests_total

  • istio_request_duration_milliseconds

  • istio_response_bytes

  • istio_request_bytes

6: Kiali in the Control Plane should now be fully functional with the Data Plane

Contributing

First, check the Contribute section in our web site, which provides a brief introduction on contributing, how to report issues and request features, and how to reach us.

If you would like to make code contributions, please also check the Contribution Guide as a starting point.