Here's a cheat tip. If your "observability tool" mentions Prometheus anywhere, it's definitely not one, it's a monitoring tool.
Similarly, if your "observability engineering" team runs Prometheus, it's a monitoring team.
I'm not saying this to be a dick. We need monitoring.
Similarly, if your "observability engineering" team runs Prometheus, it's a monitoring team.
I'm not saying this to be a dick. We need monitoring.
And I think Prometheus is a good tool, probably the last great monitoring project we'll ever need.
But it is structurally anathema to observability. Incompatible. At nearly every place where you must make a design decision, it went the opposite direction from observability.
But it is structurally anathema to observability. Incompatible. At nearly every place where you must make a design decision, it went the opposite direction from observability.
I'm not going to go into the nitty gritty just because I have talked about this A LOT. Check out thenewstack.io/observability-… or honeycomb.io/blog/so-you-wa… if you want to understand why Prometheus and o11y are like oil and water.
This has been on my mind since last week when I heard from Liz that @gitlab has an ambitious observability roadmap.
Dude, I was *stoked*. I can't wait to compete on the grounds of opinionated features and product design, instead of telling people why metrics suck all day.
Dude, I was *stoked*. I can't wait to compete on the grounds of opinionated features and product design, instead of telling people why metrics suck all day.
Then I went and read the plan.
✔️ Prometheus
✔️ three pillars
✔️ haphazard, incorrect use of terms
In other words, it's a monitoring tool with some kludgey bits. Disappoint. 😔
✔️ Prometheus
✔️ three pillars
✔️ haphazard, incorrect use of terms
In other words, it's a monitoring tool with some kludgey bits. Disappoint. 😔
Oh, link: about.gitlab.com/direction/moni…
