Been a big fan of @powersync_ for a long time! Deeply researched alternatives at the time, and have kept a close eye on the space ever since.
I haven't seen anything that does quite as much, while staying relatively unopinionated and open ended.
So much of day to day building comes down to moving data from one place to another, and a sync engine simplifies so much!
I don't know if this is a new feature or I'm just noticing it now, but thanks Anthropic.
Makes it so I can kick off some long running sessions at night and continue in the morning, without having to worry about burning a million uncached tokens.
@RobSwish@1Password Had the exact same thought yesterday while I was trying to log into X on safari on golden gate. Terrible UX, but I don’t know how much of that is safari and how much 1P.
Got a chance to go real low level and optimize performance on @TitanGymApp this weekend.
The result: actions across the board are ~2.9x as fast as before.
Had a scary almost 40min period where I couldn't even ssh into the EC2 instance that powers @TitanGymApp
Root cause turned out to be exceeding provisioned IOPS & throughput on the EBS volume.
Some thoughts:
- Sentry's uptime monitor is really useful
- AWS limits how many EBS changes you can make during a time window (I was able to quickly increase IOPS, but it didn't let me provision more throughput for another 6 hours. I probably could've batched both these changes instead of trying it one by one.)
- Despite this downtime, I still have a 99.895% availability L30D
- Because of the underlying sync engine based architecture on the app, all existing users would've experienced 0 interruptions (new users would've been unable to sign up though)
A typical json file containing geolocation data for an outdoor workout on @TitanGymApp:
- uncompressed: ~725KB
- gzipped: ~99KB
Happy to have caught this early before my s3 costs go up.