Migrate: Identifying Issues and Roadblocks
The detection of errors or possible issues with your application’s architecture, scale, or performance early in a migration can save you a very large headache. Luckily, identifying issues and roadblocks is New Relic’s bread and butter.
5 Steps to pinpoint issues and roadblocks
- Identify components of the application that can be instrumented
- Identify business KPI
- Deploy and configure products including alerting
- Setup integrations to cloud services
- Monitor for issues and roadblocks
Below is an example dashboard of what is looks like when instances increased but did not decrease. This may indicate an issue with an auto scaler:
Early detection of issues and roadblocks will allow for a quicker and smoother migration as you move to a cloud based architecture. Uncovering and solving issues early will help contain costs, too.
When identifying issues and roadblocks you are looking for metrics that are highly out of scope.
- Instances with very low CPU may indicate that you are using instances that are too large
- Very long response or transaction times may indicate an architecture that is not optimized
- Cloud costs that are exceeding normal thresholds may indicate an incorrectly configured auto-scaler, as we see in the dashboard above
Possible KPIs (these are dependent on what New Relic product(s) are used)
- Availability percentage
- Average Duration
- Page Load Time
- Web transaction time
- CPU percentage
- Average Load
- Memory Used Percentage
- Cloud Spend
- Response Time
In addition to the above product KPIs, you will now have the ability to monitor cloud based metrics.
Cloud Integrations will pull in usage, errors, and inventory data for different cloud services and offer pre-built dashboards for these services.
The image below shows both Infrastructure and cloud based metrics from the integration:
With integrations, you can easily do the following:
- Monitor Usage stats: Monitor usage stats of different cloud services e.g. how many instances, volumes, load balancers, and serverless functions are used in different regions and ability zones
- Monitor Error Statistics: Monitor any errors seen in cloud services (e.g. DNS health check, serverless function invocation errors/throttles etc.)
- Alerting: You can create alerts on any of the metrics (usage/errors) pulled in by New Relic Integrations
Create custom dashboards, too! In addition to the pre-built dashboards, you can create custom dashboards (to monitor specific KPIs for any supported cloud service).