451 karma · joined October 9, 2014
For me running it is about a basic simple thing (legs, feet, shoes one in front of the other.. not much gear) that you can scale up (speed, elevation etc) or scale down (speed, intervals/rests etc) to suit your mood. You are outside experiencing the world for better or worse (sometimes it’s cold/rains outside).
I wouldn’t run on a treadmill either it’s a terrible experience, bad for your running form too.
Now we’ve got a pipeline being defined by a declarative file being generated by code.
Reminds me of at least 1 github engineering blogpost where they reference Homebrew repo(s) as pathological examples, repos that cause them pain to operate effectively because they just don't 'look' like almost anything else.
Is what happens when you build your entire architecture around using someone elses free-tier / public service.
I'd have thought you could just pull maintenance-mode products out of the monorepo tree and stash them somewhere else. Let it rot by choice. Is basically what everyone else does, let you perform maintenance tasks on your schedule not other monorepo participants schedule.
Pulumi is more like Terraform but without HCL. So imagine instead of having to find-the-HCL-way-to-do-things and work within those constraints to make terraform work, you can use one of a few languages to build up resources (similar to you would in HCL) and have pulumi provision/act on them for you including managing their state.
Quite different to CDK/CloudFormation and much more like Terraform without the handcuffs of HCL. (sometimes handcuffs are useful tho..)
Still faaairly generous, good luck getting 50 sub requests done in 50ms if you had to measure against wallclock =)
Can I maintain state between inbound request->origin request->origin response->inbound response?
My understanding was you can’t and it got called 4 times with different events, workers solves this by you being responsible for making the origin request (there is a simple mode where it does it for you to) with promises to manage origin response and inbound response logic.. I thought lambda@edge didn’t work like that?
I guess it’s pretty similar to Apple acquiring the patents and then suing competitors.. but without all the ill will that directly suing companies off a weak base brings. Bit sad that it all rationally makes sense.
How likely are you to see 1ms execution times for this stuff tho? Super simple logic should be that low I guess..
They must have a very mature execution environment to be able to offer this (if 1ms execution is achievable), cf workers have a whole set to logic to average out cpu spikes so you don’t randomly hit the 50ms limit, bc their execution of your code isn’t that deterministic I suppose.
Interestingly i imagine a not insignificant amount of effort from Uber etc goes into combating drivers “bad” behaviors, the co-op model might align incentives to reduce this cost.
Nope 300k appr ! https://kubernetes.io/docs/setup/best-practices/cluster-larg...
This post is next level though, deeep optimization. All of it is just increasing the ceiling though, there is some limit on number of features Uber can offer in one app.. and eventually that limit will be reached, doesn’t sound like they are willing to accept being over the limit either. Wonder what space saving techniques are left in the box?
Top of mind things I found weak/weird: - Basically can’t do least privilege, only want a role to read messages from one pubsub queue? Nope not possible - IAM policies seem bolted on, legacy roles seem much better fit in the ecosystem, but they suck obvs. - different gcp resources have somewhere between very coarse grain to just acceptable iam operations. Might just be GCS that lets you do proper least privileges policies like you would do in AWS.
I was very surprised how bad it was, AWS IAM is some black magic shit that is deeply impressive and often taken for granted, even GCP can’t replicate.
ease of use beats moral advantage most times (almost every time)
Could you not just send a signal to the running process that it’s going to die and to hang up or save state or whatever is needed? If an Ephemeral vm ends in a bad state does anyone care?
In my mental model of the two problems to solve, Uber is directly consumer facing to Millions of humans and that alone creates so much work in engineering that doesn’t exist when you can craft an api with sub 50 ms latency consumed by thousands of other machines.. the humans create the complexity. Solving for their problems is why you need way more people.
i guess the tl;dr is, mature eng org shouldn't have too many dark corners for mushrooms that are important to grow, and nginx+lua choice or straight golang choice should be made informed either way.. as always.. tradeoffs! =)
There is always the benefit of building it up yourself and understanding it more deeply tho..