4 ms·
I joked about this with a coworker, but I do have to wonder what they actually use for monitoring internally. It would be interesting if they just have a second
by ollien 4y ago
I joked about this with a coworker, but I do have to wonder what they actually use for monitoring internally. It would be interesting if they just have a second copy of the prod stack for internal monitoring, or something.
- vageli 4y agoWe definitely used datadog to monitor internal systems while I was there many years ago. Dogfooding was a celebrated practice and I'd be surprised to learn that it's changed.
- ollien 4y agoI have to assume it's a separate stack from what is used in production?
- electric_sheep 4y agoI remember listening to Charity Majors[1] talk about this on a podcast -- in the case of Honeycomb they have a "dogfood" cluster than monitors production, and a "kibble" cluster that monitors the dogfood. Cute, but it gets the point across: watchmen for the watchmen, with each layer slightly less mission-critical than the last. [1] https://changelog.com/shipit/11#transcript-98 https://changelog.com/shipit/11#transcript-98