@swyx
I see it like bicycle gears. Client side rendering is gear 5. It's powerful but you have to do a lot of work to get started.
SSR is gear 1. You can start really fast but your top speed is very limited.
With react server components we can finally switch gears without...
Things Cloudflare Workers can do that AWS Lambda cannot afaik
Be a GRPC server
Handle websockets
Be stateful - with durable objects
Defer async work till after the response to the client
What am I missing?
Converted 28 million rows of data from csv to parquet with
@duckdb
It took about 5 minutes including writing and running the script with no prior experience.
I wasted about an hour trying to script this in node without it going oom a few nights ago
I have made the switch from React to Vue this year. After 6 years of react its been a big learning curve to switch it up.
My twitter timeline is full of react stuff, who should I follow in the Vue community
@MayaShavin
and
@stefanolaru
?
@ScribblingOn
From my experiance we often don't even value the right technical skills.
Getting rewarded for building quick dirty and messy, then moving on before shit hits the fan.
Spending time thinking through a problem... no time for that, gotta be coding furiously all the time.
@rem
1. A cleaner approach to error handling.
2. So I can use
@Cypress_io
stubbing.
3. Request cancellation support.
4. No extra step of await response.json()
5. Better API for post requests
6. Sensible defaults.
This is the biggest miss understanding I see about serverless on a regular basis
I use serverless because it handles all of these trivial things for me not because I can't write a Dockerfile in 2 minutes for my node server
Congratulations! You saved a few hours by not writing a Dockerfile and deploying a serverless function.
Now, good luck spending the next few days figuring out how to handle background jobs, caching, rate limiting, and all the other trivial things you need.
Happy 1 billion day!
Today
@baselimeHQ
hit 1 Billion events stored in
@ClickHouseDB
and we can still find a trace in all those billions of events in 100ms 🚀🚀🚀🚀🚀
Just figured out how we can replace a huge amount of infrastructure and glue code with a durable objects and a small amount of JavaScript. It's not even hard, just a mindset switch.
The future of serverless is stateful 🤯
#100DaysOfPigeon
#Day3
I couldn't get a picture yesterday because storm Brenden scared them all off. Today was a good day. I got a picture of 2 pigeons together
@AdamRackis
Also apparently 3x managers and 2x pms apparently exist
(Which I can confirm)
But apparently engineers are only cogs and are completely replaceable.
(Which is just laughable)
Building the Cloudflare integration for
@baselimehq
has been an incredible experience. Not only is Cloudflare a superb platform for builders, but so many folks from
@CloudflareDev
have given us early feedback, helped troubleshoot, and been super kind and helpful ❤️
This morning I was comparing Baselime with one of the leading vendors in the logging market and I was so surprised with how far we have come.
In so many ways we have a more polished, powerful and easy-to-use product 😲
Considering where we started the year it's amazing
Tell me you use DynamoDB without telling me you use DynamoDB. I'll go first: Spent an hour worrying about the cost of doing something and building a dashboard to track it when it scales only for it cost 20c 😅
Everyone is laughing about the CIA sabotage methods but I see this in large open source projects every day.
Trying to ship a feature today and some guy uses this shit to block it 2 years ago. Still this necessary feature is not shipped even though the code is good to go.
🤬
In my first weeks of being a serverless engineer I saw
@sheenbrisals
announce that Lego was fully serverless and Eventbridge was launched at
@ServerlessLDN
and was in awe!
Nearly 4 years later and still in awe
@michabahr
SQS simply keeps messages in b-tree -like structures ordered by the time of arrival. In case a delay is specified, the delay is simply added to the time of arrival. Since every message is independent, the scaling is quite simple: By default each queue gets ...
@johncutlefish
SAFe keeps army of people busy making sure everyone is busy rather than doing valuable work. Turns out being busy is not the goal. Woops
Serverless is awesome!!!!!!!!!!!!!!!!!!!!!!!!!!
Serverless is awesome!!!!!!!!!!!!!!!!!!!!!!!!!!
Serverless is awesome!!!!!!!!!!!!!!!!!!!!!!!!!!
Serverless is awesome!!!!!!!!!!!!!!!!!!!!!!!!!!
🙀🙀🙀🙀🙀🙀🙀
If performance matters in your Serverless system and you use a lot of DynamoDB streams and Eventbridge be careful.
Also SNS is absolutely amazing and we should not forget about it!
We will keep seeing this but the answer is not to leave Vercel completely.
Setting up Cloudflare as a reverse proxy has 3 steps and takes 5 minutes.
This lets you block suspicious traffic automatically and configure customizable rules if you are under attack
Oh great, I think my Vercel site is under attack.
62k failed requests in last 30m and growing
@vercel
@vercel_support
can you do something? all requests coming from "gobuster/3.1.0"
@kentcdodds
Other: Fetch at runtime from vault or secrets manager.
Environment variables injected from your ci pipeline are a bad place for secrets for many reasons
@flybayer
It was actually way easier than I expected. Must be created by some rogue employees that want to do crazy things like making our lives easier 😂
The lambda power tools team are slowly but surely turning aws lambda into a first-class platform to build robust and maintainable applications!
Anyone who is not paying attention should be ❤️
📣 Powertools for AWS Lambda (Python) 2.28 is out!
🔥 Request/Response data validation via types
🔥 OpenAPI & SwaggerUI for API Gateway REST/HTTP, ALB, Lambda Function URL, VPC Lattice
🫶 Inspired by FastAPI to ease migrations
and more!
#aws
#serverless
@GergelyOrosz
Do you think the recent expansion of this is driven by the number of tech companies realising they are at the extract phase of their life cycle?
@tuple
This breaks my workflow. I currently run a script to set my slack status to do not disturb if my cpu is raised above 69%. Can you please not ship this
@thdxr
Pythons use outside of tech is crazy though, like people are using it for some bonkers stuff for modelling and research in academia and government. I don't think their is an alternative for that. But for web dev I agree 😅
#100daysofpigeon
#Day1
I have so much fun watching the pigeons at work whilst waiting for builds. I will share the best picture I get of a pigeon every day for 100 days
Sometimes instead of removing tech debt, you can get caught in a trap of fake work to work around the tech debt because it seems easier.
It's often easier than you think to remove the tech debt
Great weekend! Now I am ready for social distancing, mesh network is setup so the wifi reaches everywhere! And I have treated myself to a new monitor 😎
@GeePawHill
Outside in TDD and frontend development are a match made in heaven. Especially with tools like
@Cypress_io
It makes your feedback cycles so fast and makes working within long and complicated user journeys so much easier. People don't know what they are missing out on!!! 🧙♂️🦄🧘🥰
If performance matters in your Serverless system and you use a lot of DynamoDB streams and Eventbridge be careful.
Also SNS is absolutely amazing and we should not forget about it!
On the cusp of releasing an open-source library... I think this is the closest I have come to finishing something. Better pour myself another glass of wine
I've been waiting for this day literally for months, today
@thomasankcorn
is joining the Baselime team
from a super early adopter a few months ago to now being part of the team!
this is us heading to making observability as code the new norm
Twitter developers: I’ve spoken at events all around the world and been interviewed for many podcasts.
Me: I write typescript in my bedroom, often wearing pants