TigerGraph Production Rollout Guidelines

INTRODUCTION

This document is for users who are starting or already started the process to develop their TigerGraph production system. It provides guidelines on the processes used by TigerGraph for ongoing engagement, implementation, and deployment. By following these recommended processes and best practices, users can better utilize the resources and support from the TigerGraph team, as well as have shorter development cycles, more stable environments, and better performance-optimized architecture. 

1. Engagement Process

The success of your project is important to all of us at TigerGraph. We will be with you through the entire process of getting into production. To ensure the smooth development cycles, it is important for our teams to communicate and stay updated. To this purpose, it is recommended to take the questionnaire as well as having kick-off and weekly meetings.

  • Kick-off Meeting

    Every new engagement starts with introducing our teams to one another, with their respective roles and responsibilities. In this first meeting, we will inform you who will be your primary contacts during production rollout and how to reach TigerGraph for your daily and long term needs. Details about your licenses, support process levels, and more, will be shared.  Another major task of the Kick-off meeting is understanding your use case. Based on the use case info collected during the kick-off meeting, we will make recommendations for your TigerGraph system configuration. For details please refer to Environment Configuration and Setup – Best Practices. Third, a timetable and plan need to be agreed upon. Based on your timeline, corresponding resources will be coordinated to support your production needs. Please refer to Professional Services for more details.

  • Weekly Meetings

    During the length of the engagement process, we recommend weekly meetings to keep everyone updated. During the weekly meeting, we will review the issues that happened in the past week, give updates on the professional service progress, as well as items that need our support during the coming week. These can be short stand-up calls to keep all the stakeholders in sync.

  • Questionnaire
    Once you have your TigerGraph environment set up, we will send you the questionnaire below in order to know more about your software version number, data scale, and environment setup. If you have more than one system, please fill out separate forms for each system. https://www.proprofs.com/survey/t/?title=qpyhg  The information collected by the questionnaire will empower the TigerGraph customer support team and professional service processes. In this way, we will always know the context of the work.

  • Hand-Off Meeting
    Once you have a running production environment we will conduct a last hand-off meeting to conclude the production rollout engagement. This call involves stakeholders, achievements, feedback sessions, and any future expectations.
  • Post Development: Health Check Events
    Depending on the contract, customers are authorized to utilize TigerGraph resources for “HealthCheck Events”. During a health check meeting (e.g., six months after the deployment) we may check your system’s speed, scale, logs, sizing and other relevant parameters to ensure continuous high performance as achieved initially.  

2. Environment Configuration and Setup - Best Practices

Having the right environment setup and server configuration is very important, since it is crucial to have enough hardware resources to support your graph analytic or real-time workload. We will make recommendations based on the available information, but the environment and configuration are the customer’s decision and responsibility. Here are some of the issues that are addressed in consultation: 

  • Ensuring your environments have the recommended hardware resources and have the right setup.
  • Separation of ad-hoc and research queries from your production traffic to avoid performing any untested operations in your production environment.
  • One of the advantages is to support both OLTP and OLAP in one cluster. When both types of workloads need to be included, we want to make sure you have sufficient hardware resources and a proper workload balance. Please contact our service team for an appropriate hardware configuration.
  • TigerGraph supports real-time graph updates. However, running a scheduled data ingestion is more performant and memory-efficient than real-time graph updates. If a real time graph update is necessary, it is recommended to consult the service team to make sure enough memory is provisioned or use a Kafka Loader or buffer the new data and send it to the TigerGraph cluster in larger batches.

3. Customer Training and Certification

It is strongly recommended to take our training and certification programs before the development. The training program covers a full series of topics from introduction to graph concepts to very in-depth query writing best practices and graph algorithms use cases. The content and duration are customizable based on the customer’s availability and needs. Training covering all training topics takes two full days, which can be split into multiple sessions and customized with only selected topics. By taking the training program, the customer team will develop all-around competence of the TigerGraph system, and be ready to take on the challenges of the development task. 

Training Objective

After the training attendees will:

  1. Gain deep understandings of the fundamentals of the graph database.
  2. Know how to operate and maintain the TigerGraph database.
  3. Be familiar with GraphStudio features.
  4. Be able to write high-performance and high difficulty GSQL queries.
  5. Become knowledgeable in solving graph problems.

Training can also be tailored to your specific needs. An infrastructure group will need a different training material from the data science team. We will be discussing your needs as the license contract. Training can both be delivered in person, or over the internet. 

Please reach out to your technical contact at TigerGraph for more details. We can provide customized training upon request.

Certification
Certification is another great approach to get your team familiarized with the fundamentals of the graph database and TigerGraph. For more information please visit https://www.tigergraph.com/certification/.

4. Professional Services

It is recommended that you utilize professional service to support your development. TigerGraph solution architects can achieve your development goals in a very short period of time and are able to solve the most difficult type of graph problems. Also, professional services make sure that your development follows the best practices. The professional service includes but not limited to the items below. 

  • Design
    We provide optimal graph schema design, solution design, and architectural design.
  • Code Review
    Review the schema design, server config, cluster setup, and GSQL queries to ensure they follow the query writing best practices.
  • Performance Optimization
    Optimize your GSQL queries to meet your performance requirement.
  • Solution Implementation
    Implement, test, and deliver your TigerGraph solution.
  • Environment Setup
    Setup the environment, optimize the system configuration. Data pipeline integration.
  • Roll-outs
    For our professional services customers we make sure that we stand by you during patches, version upgrades, and QA to Production roll-outs. 

5. Customer Support Best Practices

Customer support helps you to resolve and get around with any product-related issues. In order to efficiently utilize the support service and get the problem solved ASAP please follow the guidelines below.

  • It is okay to reach out to our support agent personally for a faster response, however, please make sure a ticket is created from Freshdesk. This way, we can keep tracking this issue.
  • If every ticket is an urgent ticket then there is no urgent ticket. Please select the urgency carefully for better resource utilization. For urgency definition please refer to our support policy.
  • Please leave as much information and description of the issue as you could, since it will reduce the iterations of communication. 
  • Please refer to the troubleshooting guide and solution pages for self-help before your ticket is responded to. Also, you can also learn how to check the logs in order to collect useful information to put in the description of the ticket.
  • Please remember to put version info in the ticket, so that in the case of a known issue it can immediately be identified by us.