Track 2: Platform runtime configuration

Abstract

This track covers the fundamentals of platform configuration when deployment of punch framework components have succeeded.

Import Operator configurations

In the previous track, we deployed a full Punch on targets VMs. If you log into your operator node, you can see that there are no Punch configurations imported such as channels, resources, dashboards ..

This operation is a post-deployment command :

source activate.sh
punchplatform-deployer.sh -cp -u vagrant

This command will copy all files located in your punchplatform configuration directory on your deployer to your operator nodes

Tip

Thanks to Punchbox tool some default Punch configurations are generated. You can specify your own using the --punch-user-config punchbox argument

Import Elasticsearch templates

Once your platform is deployed and your punchplatform configurations imported, you have to setup your platform before starting your business applications such as channels and punchlines

Firtsly, you have to import your Elasticsearch mappings, refer to this documentation to have more details.

To achieve this, log in to your operator node and execute :

punchplatform-push-es-templates.sh --directory resources/elasticsearch/templates/platform --url http://$elasticsearch_server:9200 --verbose

Tip

Replace $elasticsearch_server with the hostname of your elasticsearch server.

To check that your mappings have been successfully imported, simply run :

curl -s $elasticsearch_server:9200/_cat/templates

Import Kibana dashboards

You have now imported your elasticsearch mappings. Next step is to import your Kibana dashboards, refer to this documentation to have more details.

To achieve this, log in to your operator node and execute :

punchplatform-setup-kibana.sh --import --url http://$kibana_server:5601

Tip

Replace $kibana_server with the hostname of your kibana server.

tenants and applications deployment and starting

Your punch platform framework is now ready to go.

You can launch your business channels on your deployed platform.

However, keep in mind that your platform does not only deal with business data but also with a lot of usefull metrics.

Refer to this documentation to understand where and how data are classified in your Elasticsearch cluster

So we always want some platform-management/monitoring applications to run.

On the punchbox, you can start the example 'platform tenant' standard channels from the operator environment:

channelctl -t platform start

Then you can go to the punchbox kibana (server1:5601) and display the [PTF / Channels Monitoring] Shiva tasks events log standard dashboard. You should now be able to see the captured logs of shiva cluster and associated events.

For more detail about typical 'platform' tenants, please refer to Reference architecture

Similarily, if the integration team has designed initial pipelines for your platform monitoring, or for handling business-related data, these channels can be submitted at the end of the deployment process to begin delivering the actual business value of the platform to the customer.