To detect network anomalies, outages, packet loss, blocks, floods, disconnections, link statuses, power supply failures, UPS failures/low batteries, etc.
Basically anything that can and will go wrong with the network, Spectrum will give you the drop before user community is aware.
[Sentry] is honestly an amazing product. It allows us to detect errors in real time complete with stack traces and any extra accompanying information the developer wants to provide in the alert. With the alerting into Slack it has allowed us to quickly triage and tag in people who need eyes on a specific issue. It would be really useful in any Saas product environment.
To be honest, Spectrum is one of the most stable and less problematic solutions that we work with. It’s easy to scale and to deploy. The only time we’ve had problems with it was with some upgrading processes. It is not so intuitive and you have to make sure everything is correct.
We actually ended up using both because New Relic is a more robust overall IT infrastructure monitoring product. However, sentry is more developer oriented on the backend and more client friendly on the front end as far as showing results and the dashboard etc. It can provide product level insights that New Relic does not.
Error tracking is a must in any modern dynamic website or app. By looking into the error notifications I'm able to fix errors before anyone even has a chance to complain about them!
Surprisingly, many website issues aren't showing up in Sentry, because they don't trigger exceptions. I'm interested in seeing if I can use Sentry to catch manually-triggered exceptions for "undesirable states" that my website can find itself in. Of course, that means I have to figure out how to have my client code recognize that it's in an undesirable state...
