Prepare for high volume events¶
Use the following considerations and recommendations to plan for high seasonal traffic. Your Rackspace Account team and RAS Digital Experience team can work with you to understand and implement these recommendations. They come up with a Seasonal Readiness Plan for your environment to ensure that your end users are not impacted by poor performance during peak seasonal traffic periods.
Limit code deployments leading up to peak traffic periods to minimize changes in the environment.
Consider limiting or disabling back-end administrative functionality, such as imports, catalog updates, and so on, during peak traffic periods.
Consider scheduling regular, short touchpoints with all appropriate parties throughout the peak period.
Two to four months before peak traffic begins¶
Perform load testing: Engage the Rackspace Customer Success team, RAS, and Network Security to participate in load testing to observe, analyze, and take action on the results.
Depending on your environment’s size and complexity, use marketing analytics data to calculate the expected peak flow of traffic and perform load testing. Refer to the following guidelines when you plan for and execute application load testing:
Many environments can expect a 20-30 percent increase in traffic year over year.
Develop a load testing plan that tests all application functions and tests for at least 125 percent of expected peak traffic.
Use a load-testing tool that incorporates realistic think time to gauge the impact of real users more acurately. If you do not have a preferred vendor or tool, Rackspace has load testing partners with whom you can engage.
Test against an environment that uses the same codebase and is as close to production as possible. Modern, complex applications do not scale linearly, and you cannot always accurately extrapolate real-world performance.
Use metrics to establish the success or failure of a test. For example, a test is successful when at least 95 percent of transactions complete without errors.
You might need to pconduct multiple load tests to ensure that you can iterate through the results and achieve stability at peak user traffic levels.
One to two months before the peak traffic period¶
Integration points review: Work with the RAS team to review integration points approximately one to two months before the peak traffic period. Include internal and external integrations such as search, payment gateways, OMS, inventory, mobile, and so on.
Ensure that load testing includes integration points.
Where possible, review integration point architectures to determine if they are highly available.
Understand the impact on users if the integration service is degraded or unavailable.
Configuration review: Work with the RAS team to proactively review your environment’s configuration and APM metrics, including the application, web servers, caching layers, databases, and so on. This review can uncover bottlenecks that developed during load testing.
Ensure that you allow for time to make changes and then test again.
APM configuration review: During the APM configuration review, engage with the Rackspace team to:
Review business transactions configurations
Ensure that you configure health rules appropriately for the environment
Review dashboards to reduce troubleshooting time in the event of issues
Two to four weeks before the peak traffic period¶
Network layer review: Engage the Rackspace Account team and Network Security team to review performance of firewalls, load balancers, and any inline devices, such as IDS, to ensure that they can handle increased traffic.
CDN and caching tuning: Engage the Rackspace Account team and RAS to ensure as many static assets as possible are cached. Ideally, the CDN or cache layer should take as much traffic as possible, which reduces calls back to the application.
One to three weeks before the peak traffic period¶
Schedule jobs review: Engage the RAS team to review any scheduled jobs, such as cron jobs, scheduled tasks, and backups, to minimize impact during peak hours.
Review URL monitoring: Engage Rackspace RAS to ensure synthetic transactions are in place to monitor the availability and functionality of critical business paths (for example, add to cart, checkout, search, and so on).
Review account guidelines: Engage the Rackspace Account team to verify that escalation paths, return-to-service instructions, and runbooks are up-to-date and technically valid.