Edge computing – with its ability to help unleash the potential of smart cities and autonomous vehicles, is something often thought to exist “around the corner” as opposed to an industry already in motion. In Gartner’s 2020 Market Guide for Global CDN, the research agency actively defines CDNs as “highly distributed edge-based cloud delivery platforms” and argues that CDNs must provide capabilities in edge advanced services, such as supporting microservices and enabling local and global edge code execution. Many of the CDN providers we work with at Catchpoint, in addition to various enterprises, are already providing scalable solutions at the edge and are redefining themselves as edge platforms.
In this post, let’s take a look at how you can leverage Catchpoint’s services, whatever stage of the journey you’re at, to monitor the edge today. More specifically, we’ll take a look at three of our edge-specific capabilities:
- Custom Metrics and Customized Data Breakdown
- Request Override
- Network Insights
Together, these capabilities provide next-gen visibility as you can troubleshoot edge points of presence (PoPs) and servers.
Isolate Performance with Custom Metrics and Customized Data Breakdown
Not every monitoring tool offers the opportunity to isolate performance based on specific dimensions and to break down data in a highly detailed way. Indeed, your traditional monitoring platform may allow you to look at overall response times or in some cases, allow you to break this data down by cities, as below:
However, when looking at the end-user experience in the context of edge performance, you need another level of granularity. You need the ability to breakdown data from each city by the edge server serving the request.
With Catchpoint, you can monitor edge Points of Presence (PoPs) and/or servers using custom metrics and breakdown data at a granular level. When you create your own breakdown, you can analyze the data more effectively. Having access to detailed, granular telemetry can help DevOps and IT teams significantly reduce MTTR as they can see exactly where a problem lies and quickly take action to solve the problem.
To get started, simply hook Catchpoint’s ability to capture custom metrics into any data that is present in your HTTP response header or HTTP response content. In the response header, most edge solutions include the name of the server serving the request. Once the server name is extracted, you can immediately examine the performance of each of the edge locations. That’s the level of detail possible with this feature. It gives you the ability to determine which edge servers are optimal for a given end user location.
What can you do with this context? If you have serverless application logic running on a CDN or edge network and you have users coming from multiple locations across the globe, you’ll want to know which edge server they’re hitting so that you can quickly ascertain how that particular server is performing from the end user perspective. If you have users in multiple locations along the West Coast, for instance, you can ask if it’s logical to have your application logic served out of only one edge location, such as Portland or LA, or if it makes sense to have it more widely distributed. You can then use that information to decide where to serve a particular request to maintain optimal performance for that end user.
Using this capability, you can also make decisions about whether to direct all the traffic in a particular region to a specific PoP. Some edge platforms and CDNs choose to have more capacity on one PoP location than another, for instance.
The ability to make these kinds of detail-led decisions is why it is helpful to breakdown data at a more granular level so that you have the advantage of understanding the performance of each of your edge locations from the end user’s perspective.
Verify Requests with Request Override
From a provider standpoint, it’s important to keep a tab on every PoP where you’re providing this kind of functionality. And from a developer perspective, you need to ensure that the right request is fetched for users in different countries. This is where the concept of Request Overrides comes into play.
Analyzing the performance of specific CDN PoP or edge, or ensuring the PoPs are working as expected is a challenge in production. With the request override, you can test and compare performance without making any changes to the production environment. The image above illustrates how a request override is set up in Catchpoint. Whenever a user makes a specific request, for example, www.example.com/edgelogic.js, it will resolve to one of the various PoPs or edge locations that are present in the CDN or edge network infrastructure. Instead of needing to make any kind of production change to point to a specific edge, say www.edge01.exmaple.com, you can simply type, www.edge01.example.com/edgelogic.js in the Value field, allowing you to immediately perform analysis on a new edge location. You can also target a specific IP address using this feature.
What you’re doing is simply redirecting the traffic from one location to another. This allows you to perform monitoring of each and every edge location that you have from a vendor perspective.
Request Override is an extremely powerful feature and many of our CDN and edge compute customers use this regularly.
Utilize Network Insights to Gain Visibility into Every Network Layer
Catchpoint is the only monitoring vendor to offer synthetic monitoring of the full stack. Network Insights offers a suite of capabilities to diagnose complex performance issues across network monitoring protocols, including Traceroute Monitoring and DNS Traversal Testing, both of which are particularly helpful for customers with global footprints and complex architectures and understanding performance from the point of view of the end user, regardless of where they are in the world.
Verify the Paths to Your Edge Locations are Working as Expected
Catchpoint provides two different types of Traceroute visualization: a logical Autonomous System (AS) view and an IP/hop-by-hop view. Each one shows the health of each router in the network path and can be run on three network layer protocols: TCP, ICMP, and UDP. Traceroute helps you understand if the paths to a specific edge location are working as expected, performance and reachability wise. By selecting the IP network view in the Catchpoint smartboard, you can utilize Traceroute to plot out the different traffic patterns to each PoP. This will let you see any packet loss and routes taken, including their status all the way to the destination address at the edge.
The Traceroute view shows you precisely which hops are causing packet loss and understand specifically which providers are experiencing problems when. You can share this detailed data, in order of issue escalation, directly with the network provider to get end user performance back in short order.
Verify and Update Mapping Decision Criteria with DNS Monitoring
Monitoring DNS, meanwhile, allows you to verify and continuously update decision criteria and ensure it is working as intended. The load balancer, which determines the handing out of edge servers, is often determined at the DNS layer, meaning this is the layer at which many critical decisions are made, including specifically which edge will be used.
Based on this data, you can determine where the user is directed, for instance, if a user is in central California, you can determine if the optimal edge location is LA or San Jose.
Monitor the Edge with Catchpoint
Catchpoint offers several industry-leading solutions to monitor the edge today, allowing your NOC and DevOps teams the ability to make optimal mapping decisions specific to the needs of your end users, regardless of where they’re located.
To learn more about Catchpoint’s CDN and edge monitoring solutions, watch our CDN Monitoring webinar.