Due to the nature of cron jobs, our ping endpoints (https://hc-ping.com/... URLs) receive spiky traffic:
On average we write ~500 ping requests/second to the database.
At the start of every minute, we write around 4000 pings/second.
At the start of every hour, we write over 10'000 pings/se
Interesting read, it does seem like time to start looking at horizontally scaling the workers based on request pressure. Having a dynamic number of workers that can do the database updates and then quickly release the mutex should increase throughput.
Interesting read, it does seem like time to start looking at horizontally scaling the workers based on request pressure. Having a dynamic number of workers that can do the database updates and then quickly release the mutex should increase throughput.