Charity Majors @mipsytipsy CTO @honeycombio, ex-Parse, Facebook, Linden Lab; cowrote Database Reliability Engineering; loves whiskey, rainbows. I test in production and so do you. 🌈🖤 Dec. 10, 2018 2 min read

In honor of NewRelic sales team using my critical tweets about Datadog's host-based pricing scheme to try selling their own shit, here are some tweets that Datadog might like to borrow.

First: pretty ballsy, considering that NR themselves only just switched to volume based pricing recently. How recently? I'm not sure, but renewing customers aren't being migrated on to it, and I hear of other shenanigans. (don't know the magic words? you pay the sucker price)

Sadly, it seems a generation of engineers have learned to optimize infrastructure for their monitoring service provider's pricing schemes, not for distributed systems operability or reliability.

Second, the increases. 2x-5x seems common after the first year. Pretty shady, since they're banking on the fact that your team has gotten used to using the tool and nobody has time to learn new tools to make you cough up.

Third, the "you must deploy to 100% of nodes and keep 100% of events" boondoggle. O, let me count the ways that this is wrong -- and rest assured, this is a revenue strategy from start to finish, not a technical reason.

*Most of the things that happen on your system are boring*

And NewRelic has all the information it needs to increment a counter when it sees the same boring event again, instead of storing it and charging you.

For more on how real engineers approach dynamic sampling, see about halfway down @dotpem slides here :  https://qconsf.com/system/files/presentation-slides/nathan_leclaire_-_qcon_sf_2018_-_observability_the_health_of_every_request_.pdf 

This whole "you must keep every single event" bullshit is proof that engineers can be just as vulnerable to sales hype and marketing campaigns as your average teen shopping at forever21. 💅

Do you know how many tons of cash you have burned on the pyre of this lie? Be angry.

HALT, I STAND CORRECTED. Their "cloud based" pricing model is literally just per (CPU cores+GB ram) * hours used.

And browser is per page load. And... tbh there seems to be a different pricing scheme for every which way you might want to store or access your data.

I actually can't tell how or why they think they can throw shade on datadog for pricing reasons, they seem like two malformed heads on the same radioactive puppy to me.

Alright, I don't like kicking someone when they are flailing on the floor, so I will stop there.

Interesting to note the ways pricing gets distorted by sunk costs and ongoing revenue, not what's best for the user/what aligns incentives. But saving my theories for another thread

Oh but p.s. You can do dumb sampling perfectly well, just run NR on 10% of of your nodes. (As nearly every person who has ever used NR has figured out.)


You can follow @mipsytipsy.



Bookmark

____
Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Threader is an independent project created by only two developers. The site gets 500,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Running this space is expensive and time consuming. If you find Threader useful, please consider supporting us to make it a sustainable project.