Internet Sonar leverages the world’s largest independent active-observability network to deliver simple, trustworthy internet health information, helping you stay ahead of productivity- or experience-impacting incidents. It monitors hundreds of popular services across key categories such as internet infrastructure, SaaS, and MarTech, providing real-time updates through an interactive map, dashboard widget, and API. Internet Sonar helps you ensure a resilient Internet across all these dependencies and deliver exceptional digital experiences for your customers to drive business success. If you don't already have access, you can enable services a-la-carte using your existing points. Each enabled service consumes only a few points per day, so it’s easy and affordable to give Internet Sonar a try.
Incidents detected by Internet Sonar can be correlated with downtime and errors found in any tests you are running. For example, you may be testing your website "helloworld.com" which runs on AWS and uses Akamai CDN. If Catchpoint determines that an outage at AWS or Akamai has caused downtime for helloworld.com, this will be indicated in various analysis pages like Smartboard and Records.
Here is the list of services monitored by Sonar.
Enabling Sonar
Internet Sonar is accessed either by purchasing a contract which includes specific services or service categories, or by using points. When using points to access Sonar, you can enable any individual service for any given timeframe. Each services consumes points on a daily basis, allowing you to turn on and off any individual service at any time. You only use points for the number of services active per day.
By default, Sonar is included with your account including 1 complementary service. You can change that service at any time in control center by deactivating the existing free service and activating any other service. Any additional services added beyond the 1st free service consume a small number of points per day while they are active.
If you don't have an Internet Sonar contract or any licensed services, then by default the Sonar Overview dashboard reports anonymized incidents for services that haven't been enabled with points. For example, if a CDN service has an incident and you haven't activated that service, Sonar will report generically that a "CDN Service" has an incident.
Clicking the incident will give you the option to add the service. It will then display which service had the incident, and that service will begin to consume points on a daily basis.
You can also add services from Stack Maps or from within Control Center under Internet Sonar > Services.
If your account previously used a prepaid contract that expired, Stack Map will display the following message:
You can reactivate any Sonar services within the Stack Map. Doing so will ensure the services report incident data from Sonar and will consume points.
How Service Outages are Detected
Internet Sonar detects service outages by analyzing test results for the specified services. A service outage is typically identified when any test or request to the service has >20% downtime for a 15-minute interval, or the service has a trend shift in downtime. If a test or request is failing consistently for a given node, this will be represented as an outage for the city where the node is located. As outages are associated with the cities where the affected node is located, outage data in pages where Collections are required will only be shown if the city where the outage is detected is located in the region purchased as part of the Collection. For example, if Catchpoint detects a service outage in NYC, it is reported under the North America region. Thus, if you purchased a Collection for that service in North America, it will be reported in the Overview and Custom Dashboard.
Service Outage data is collected from two sources as described in the following sections.
Tests Managed by Catchpoint
These tests run from the majority of backbone nodes, with each test strategically targeting specific nodes based on the monitored service. For example, AWS regions are tested by nodes close to each respective region. Certain social media platforms are not monitored from China where tests would always fail due to firewall policies. The tests vary in frequency, but most run at five-minute intervals. Critical services are monitored using multiple tests for different functions and protocols, and some services are inherently subject to redundant monitoring; for example, a CDN might be monitored directly with one test, and also indirectly via another monitored service that uses the CDN, such as a SaaS application.
Tests Run by Customers
Some customers have opted to allow their test data to be anonymized and used for outage detection. We do not collect or report any personal- or company-identifying information related to these tests, nor anything else that could be considered private or sensitive. The collected data simply consists of downtime and error data for public 3rd-party services found in tests.
How Network Outages are Detected
Internet Sonar detects network outages at the Autonomous System (AS) level using traceroute tests running on all public nodes across our platform covering tests to various internet services and infrastructure like DNS, CDNs, clouds, and ISPs. If the traceroute test has packet loss and the destination ping fails, this is taken to indicate a network outage. This process eliminates false positives in cases where packet loss might not be relevant to the route, like routers that drop ICMP packets.
Sonar also captures and reports predecessor hops (the hops preceding the failed hop in the traceroute.) This helps with understanding exactly where in a route the packets are being dropped, e.g. between two routers within the same AS versus two routers from different ASes. In some cases, a traceroute hop might result in no response at all, meaning we experience 100% packet loss. When this occurs, we get no IP address for that hop, and thus the AS and geo-ip info is unavailable. The predecessor hop data is vital in these cases as it provides the last known AS and IP data available from the tests.
Network outage severity is based on the number of test runs with packet loss for a given AS.
Outage Analytics
Internet Sonar data is reported in multiple locations within the Catchpoint Portal as listed below. The main data reported includes the name of the service experiencing an outage, the region(s) impacted, and outage start/end times.
Incident Severity Score
Incidents are assigned a severity level of "low", "medium", or "high" based on a combination of factors, including how widespread the outage is (number of geographic regions affected), the total % downtime detected, number of failed test-runs, and the number of IPs affected.
Internet Sonar Overview Dashboard
The Overview Dashboard includes a tab for Internet Sonar which reports outages for any purchased Collections. Each service outage is displayed with a list of error types we've observed, which may include: DNS, Connect, SSL, No Response, Response Code (400/500), and Timeout. You can also click on a an outage to view full details.
Custom Dashboard
Custom Dashboards allows you to create dashboard widgets for the services and regions included in any purchased Collections.
Custom dashboard widgets report Sonar incidents using Incident score as a range from 0-100 where 0-33 indicates a low severity, 34-66 indicates medium severity, and 67-100 indicates high severity.
Smartboard
With the base contract (independent of any purchased Collections), the Smartboard for Web/Transaction tests displays details about any regional service outages that correlate to an outage in the specific test you are viewing.
Records
With the base contract (independent of any purchased Collections), Records for Web/Transaction tests display details about any regional service outages that correlate to an outage in the specific test you are viewing.
Note that the services you use have their own service dependencies, and outages in these other services may also be correlated with your tests and result in alerts from Internet Sonar. For example, Google Analytics runs on Google Cloud, so Google Analytics users might see correlations for Google Cloud outages in their test results. If Internet Sonar correlates outages for services you don't use directly, you can check the waterfall requests and run whois on the IP for a domain in question to see who owns the IP for that request.
Alerts
Standard Test Alerts with Correlated Outages
With the base contract (independent of any purchased Collections), standard alerts for Web/Transaction tests display details about any regional service outages that correlate to a failure in the specific test generating the alert.
Outage-specific Alerts
You can also define alerts specifically to notify you of service outages. These alerts are configured in Control Center under the Internet Sonar section, and can be triggered by outages in all services and regions, or filtered to specific service/region combinations. You have the option to alert on any outages to the respective services/regions, or you can choose to alert only for outages affecting specific tests/products.
