We use Redis for buffering for a certain timeperiod, and then we write data for one sensor for that period to S3. However we fill up large Redis clusters pretty fast, so we can only buffer for a shortish period.
We tried some readymade options but they were way more expensive than our custom built S3 solution (by a factor of x10 approximately). I think we tried timescale and AWS Timestream. I haven't heard of SeaweedFS.
The problem is that the initial writing is already so expensive, I guess we'd have to write multiple sensors into the same file instead of having one file per sensor per interval. I'll look into parquet access options, if we could write 10k sensors into one file but still read a single sensor from that file that could work.
Does anyone have a good alternative for storing large amounts of very small files that need to be individually queriable? We are dealing with a large amount of sensor readings that we need to be able to query on a per sensor basis and a timespan, and we are dealing with the problem mentioned in the article, that storing millions of small files in S3 is expensive.
As far as I remember, Apple was planning to detect CSAM, and upload the detected images to the cloud, showing the photos to apple employees. This is not at all comparable to a locally running API that returns something like "{nudity: true}".
I think the author of the article got a few things wrong. As far as I understood, SafetyCore is not "scanning all your photos". It is an API that apps on your phone CAN use to LOCALLY detect nudity and other potentially inappropriate things in photos.
If I understood the article correctly, the high safety comes from having two ground connections and stopping already if one of them is broken. Your way would mean that if PE breaks, and at the same time contact between some high voltage component and the chassis is made, electrocution is possible.
A cheap car costs around 300€ per month (including taxes, repairs, loss of value, etc.), or way more if you drive a lot or if you have a car that is new-ish. I don't think that 4-8 deliveries per month would cost more than 80€ per month in delivery cost. Also, what's keeping you from going shopping by bike? We do it all the time.
How unlimited is it, really? How much do people take off in reality? I dislike these "benefits" that no one then dares to use in reality. We have 6 weeks off, and you are literally forced to take them. This basically prevents social norms from putting pressure on people to not go on holidays as much or not during certain times.
My partner and I are in the top 20-25% of Germany (looking at combined household income, we earn about the same). Not poor, but not (very) rich as well.
Super cool, I just used your app to figure out where to place my clothes drying rack so it'll get sun sooner! Okay, I already new the result mostly, but still fun and useful!
I'm European too, and I also usually don't spend more than 2000€ on holidays per year, but when I compare myself to friends and colleagues, I always had the impression that how we go on holidays (camping, traveling by bicycle/train) was relatively uncommon. Therefore I didn't think that 80% of the population went on holidays like that.
100$ for four beds including breakfast? I live in Europe, perhaps that's why the prices feel off. I pay around 100$ per night without breakfast here in most places, for 2 people.
I was very surprised by the holiday budget. Less than 2500$ for 80% of the population? Does that mean that no one from these 80% stays in a Hotel or AirBnB for longer than a week with their family? No one travels overseas, or if so, only every 5 years or so? 2000$ is what you can easily spend on plane tickets for a family of 4.
I have an ebike and a trailer, and it absolutely removed my need for a car. My neighbors are a family of 6, and an electric cargo bike allows them to live without a second car.
You must have gotten a weird Garmin that doesn't suit your needs then. Why did you get a Garmin with a touch screen, when you really wanted buttons and Garmin offers watches with buttons?
And I believe the GPS sync is necessary when you don't have an internet connection on the watch.
My Garmin Forerunner has everything the author claims doesn't exist except the hackability: Long (10 day) battery life, reflective screen (plus background light), physical buttons, all the features (step count, notifications, music control, weather, etc. plus lots of additional sport functions and even a pulse and oxygen sensor.
However, I still get the point of having a hackable device.
While it sure looks like fun, it doesn't really seem useful. All the use cases the author makes are more gimmicks than anything really useful. Actually, the thing influencing compasses seems really annoying. You'll never be able to have your smartphone show the correct walking direction while holding it in the hand with the magnet.
Unless it processes images on the fly to do some sort of image-based steering or target recognition, an Arduino or comparable microcontroller should be beefy enough [1].
In mobile applications, power consumption of a RPi quickly becomes an issue.