Discussion
Pegasystems Inc.
US
Last activity: 22 May 2023 10:06 EDT
Ask the Expert - Kubernetes with John Peterson
Join @JohnPeterson in this month's Ask the Expert session on Deploying & Running Pega in Kubernetes
Session Dates: 8 May - 19 May
Make sure to Follow and Favorite for updates!
This session will open for questions on 8 May!
Meet Your Expert:
John currently leads the Platform Deployments & Data Services group who provide the deployment, installation, update, maintenance, and troubleshooting automation of Infinity to Pega Cloud. His teams maintain the Pega Helm Charts and Pega Runtime and Pega Installer Docker images which may be leveraged for Client Managed Cloud solutions and are available on GitHub. John has been with Pega engineering for over 17 years and has been involved with the efforts to support Pega deployments to Kubernetes from the beginning.
Message from Your Expert:
I look forward to learning how Pega is being deployed to various client managed cloud solutions.
Pegasystems Inc.
IN
Our May Ask the Expert session is now open for your questions!
Ask the Expert Rules
- Follow the Pega Support Center's Rules of Engagement
- This is not a Live Chat - @JohnPeterson will reply to your questions over the course of the session
- Questions should be clearly and succinctly expressed
- Questions should be of interest to many others in the audience
- Have fun!
Pooja
Virtusa UK Ltd
GB
@JohnPeterson Thanks for the Q&A. I am new to containerisation and Kubernetes. Could you please advise what pega devs should be knowing regarding containers and k8s, what is the correct starting point for learning and what are the various resources(within and outside pega community) you recommend to learn to execute pega projects.
Thanks
Ganesh
Updated: 10 May 2023 15:17 EDT
Pegasystems Inc.
US
Hi Ganesh, for starters I would start by becoming familiar with Docker and how it encapsulates functionality (libraries, configuration). Once you understand that concept, you can move on to Kubernetes, and learn how it orchestrates multiple containers that work together. You should also explore Helm and understand the value it brings when it comes to deploying software to Kubernetes and rolling out changes. When finally ready to deploy Pega, you can follow the documentation that is available on Pega's Github https://github.com/pegasystems/pega-helm-charts which includes runbooks for specific Kubernetes providers that are supported by Pega.
Have fun!
Virtusa UK Ltd
GB
@JohnPeterson Do we have a plan to reduce/eliminate the reliance on database for deployments?
Thanks
Ganesh
Pegasystems Inc.
US
Hi Ganesh,
No. The database contains Pega Infinity's code, rules, data, and case items.
Bits in Glass
CA
@MarissaRogers Is there any way we can learn Deploying & Running Pega 8.8 in Kubernetes in our local machine?
Pegasystems Inc.
US
The recommended Pega Infinity 8.8 platform deployment leverages externalized Elasticsearch, Hazelcast, and Kafka. These services have their own set of CPU and memory requirements, and similar to the database, they have to exist prior to deploying Pega Infinity 8.8. Assuming these services are being made available to your local machine, you could potentially run one or two Pega pods on a local development Kubernetes cluster software, like Minikube. Keep in mind that the Pega provided Helm charts contain subcharts that will automatically set up these services on the Kubernetes cluster for you, so you'll want to explicitly disable them in the Helm chart and reference the existing deployments you plan to leverage instead, prior to your deployment of Pega, unless your local machine has the resources necessary to run everything.
Anamata
NL
@JohnPeterson @MandeepRawat I found this topic is very useful: https://support.pega.com/discussion/deploying-pega-8x-kubernetes (there is an attachment in this topic. So make sure you're logged in)
Folksam
SE
@JohnPeterson thanks for the Q&A. Our target platform is Azure Red Hat OpenShift (ARO) which isn't currently in the list of supported Kubernetes environments - is there a technical reason why OpenShift is only supported on-premises?
Thanks,
Daniel
Updated: 10 May 2023 15:17 EDT
Pegasystems Inc.
US
Hi Daniel,
I'm not aware of any technical reasons why Pega wouldn't work with the Azure Red Hat OpenShift offering, but we do not have the testing infrastructure built out to verify.
We do support Red Hat OpenShift.
Updated: 11 May 2023 6:46 EDT
Folksam
SE
@JohnPeterson thanks for the reply. To clarify, does that mean running Pega on Azure Red Hat OpenShift is supported or would we be running an unsupported configuration?
Thanks,
Daniel
Pegasystems Inc.
US
Cognizant Technology Solutions
GB
@JohnPeterson - I am new to Docker and Kubernetes. I was trying to install Pega on docker initially ( without Kubernetes as I am not looking for orchestration now). I downloaded the Pega docker images and ran the platform/installer to populate a local database (PostGres) with the tables. After that I tried to run the platform/pega image in a container by passing in the DB details. While the container is running and I can see Pega is up , how do I try to connect to it using the browser? If I do - http://localhost:8080/prweb/PRServlet it times out. I also tried the port mapping while running the container but it doesn't work. Can you please advise how to deal with this?
Thanks and Regards,
Samiran
Pegasystems Inc.
US
Hi Samiran, it sounds like you probably need to read up on your Docker container networking since it is likely your 'localhost' is referring to your laptop instead of the running container. Start with https://docs.docker.com/config/containers/container-networking/
Keep in mind that a single Pega runtime container still requires the backing services Kafka for Queue Processor item storage and SRS for Search Indexes storage in addition to the database, so don't plan on actually being able to do any Pega development with this configuration when you finally see the login page.
Other than that, this does seem like a cool learning Docker exercise! Good luck!
Cognizant Technology Solutions
GB
NCS Pte. Ltd
SG
Hi @JohnPeterson : When moving to external services, is it mandatory to use all the external services?
I know deployment of Cassandra is not required if there is no Decisioning module.
Regarding Kafka, the understanding is that it is mandatory for que processors. Do we need the stream pods to run if the kafka is externalized?
Regarding SRS, understanding is that SRS pod is mandatory even if elastic search is externalized. Is the understanding correct? If an application is not using any of the elastic search features (like case search and other reporting needs), is it mandatory to setup SRS pods and external elasticsearch?
Thanks.
Updated: 16 May 2023 13:09 EDT
Pegasystems Inc.
US
Hello Arul,
The embedded services are deprecated as of 8.8. Please see https://docs.pega.com/bundle/platform-88/page/platform/hub/platform-install-update.html for more information.
No, if you are leveraging externalized Kafka, then you should not have a stream node type.
Yes, the Search and Reporting Service container (SRS) is required as part of an externalized search deployment of Pega.
Pegasystems Inc.
US
Thank you to @JohnPeterson for being our amazing expert in our first session on Kubernetes! And thank you to everyone who took the time to ask a question and check out this session!
This session will now be closed to replies, if you have any Kubernetes questions - please ask a new question.
Keep an eye on our Ask the Expert landing page for our upcoming sessions!