ILT: OA610: Benchmarking and Capacity Planning
Leverage an open-source tool to generate synthetic workloads, analyze their behavior through external observation and internal database metrics, and adjust your hardware configuration to evaluate its impact on performance.
Advanced Workshop: Benchmarking and Capacity Planning (OA610)
Overview
If you want to provision the right infrastructure for your MongoDB cluster to satisfy your production workload without over-paying for unnecessary resources, it’s essential to understand the implications of infrastructure changes for your workload and how to measure what components of your infrastructure are under-, over-, or right-sized.
In this training workshop, we use an open-source tool to create a number of synthetic workloads, observe their behavior both externally and using internal database metrics, and modify our hardware configuration to observe the impact on performance. In doing so, we learn best practices for MongoDB benchmarking and how to interpret the monitoring metrics in Atlas / Cloud Manager / Ops Manager correctly.
This training workshop will arm you with the skills to determine when adding CPU, RAM, or disk IOPS will be necessary to improve the performance of your workload, and whether you are spending too much on one thing or not enough on another.
Intended Audience
This training workshop is intended for operations professionals who have completed foundation training to at least MDB200 (Diagnostics and Performance Tuning) and want to be able to right-size their database infrastructure.
Workshop Objectives
In this training workshop, you will learn how to:
simulate client workloads using POCDriver, an open-source MongoDB workload generator
identify which resource is currently limiting performance
determine how the performance of common workloads relates to the number of available CPU cores
quantify the impact of too little RAM and estimate required RAM for your workload
identify when disk IOPS is a bottleneck and how many is enough
identify whether a bottleneck is in the database or elsewhere and avoid wasting time looking in the wrong place for a solution
Duration
1 day
Program Datasheets
Course prerequisites
This course has the following prerequisites:
- MDB200: MongoDB Optimization and Performance
Language
This course is offered in multiple languages. Coordinate with your MongoDB representative to find out more.
Course access
This course is entirely virtual and will be conducted over Zoom video conferencing and Instruqt hands-on lab environments.
You can view all upcoming public training sessions via the MongoDB Instructor-led Training Schedule.
If you've already purchased training, and you're looking to register for an upcoming session, email training@mongodb.com with the following information:
- Public Training Session Name
- The names and email addresses of all attendees you'd like to register
If you are interested in purchasing seats for one of these upcoming sessions, speak to your MongoDB representative, or contact us directly.
This course is available as a private session for Precision Learning Programs (PLPs). With private training, you can select specific courses to deliver to your team in a dedicated classroom, on a schedule that works for you.
Speak with your MongoDB Sales representative to build a custom training plan.
When you're ready to schedule your private sessions, coordinate with your Professional Services representative to select delivery dates.
Please note: minimum requirements must be met to schedule private sessions.
