Hi, just wanted to know the bandwidth per HTTP/HTTPS request that incurs due to telemetry data being sent
It totally varies per-instance. Some just a few KB, others a few MB. Busy servers with lots of unique clients will send more data. Busy servers with the same clients over and over will send less data. Idle servers will send the least. You’ll just have to measure it and find out for your specific site.
Just to be unambiguous, Caddy doesn’t report in real-time, so there’s zero bandwidth impact on individual requests. Instead, it checks in to send averages over intervals.
The reason I’m asking, is that I want to support the project, but I’m expecting around 550M daily requests to my servers and I don’t want to incur the cost of that (AWS is pricey like that )8 )
Honestly, this sounds like a statistic that the telemetry server is well positioned to provide.
Should be possible to start keeping a record of the submitted telemetry package sizes, right? Then it could be compared with request rate to get an understanding of how they correlate.
A transparent average of how much(/little) bandwidth is likely to be consumed by enabling telemetry can’t hurt.
Well we do just dump the JSON payload into the DB, so we already basically have that info (sans HTTP header, etc) – just have to do the actual queries to get the numbers.