What are some best practices when configuring a test?
There are numerous best practices to follow when configuring a test. Three of the most critical ones are as follows:
- Pick multiple nodes to avoid a single point of failure.
- Optimize your alerting strategy to avoid false positives.
- Consider capacity restraints when you are setting up your tests.
These practices are detailed below.
Pick multiple nodes to avoid a single point of failure.
If you are running tests only on a single node, you are setting yourself up for the possibility of multiple problems. For instance:
• You are getting only a partial picture – You can only see what is happening on a single ISP, in a single location.
• You are not accounting for possible node failure – If the node you are using fails or is paused, your monitoring activities will be paused as well.
• You may get false alarms – Triggering an alarm every time a node is down can add noise that may not be helpful.
For a more comprehensive and effective monitoring strategy, best practices dictate the following (Fig 1):
• Choose multiple nodes.
• Pick several nodes in the same city.
• Pick several nodes in adjacent cities.
• Select nodes on a variety of ISPs.
It is also recommended to create node groups for the ease of adding/removing nodes across multiple tests.
Optimize your alerting strategy to avoid false alarms.
The alerting strategy should account for node-related failures. An ideal test configuration should include multiple nodes (different ISPs and locations). Triggering an alert every time a single node is down will only add to the noise. It is recommended to set up alerts only when 50% of the nodes report an issue. This will reduce unnecessary alerts and help ensure that only valid incidents are reported and acted on. Your CSM and CSE can work with you to optimize your alerting strategy.
Consider capacity restrictions when setting up tests.
Since Catchpoint is dealing with servers, rack space, and power, capacity is an important factor for us to continue to properly serve all our customers and bring them the highest possible value. Although we review our capacity thoroughly every month, customers must notify us in advance in the event of any large-scale testing need that they may have. This helps us avoid overwhelming our infrastructure and prepare and manage the increased load effectively, without impacting other customers.
Please raise a ticket with our customer support team (support@catchpoint.com) to address any such requirement.