Rendered at 22:32:48 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
varun_ch 48 minutes ago [-]
I wonder what people think about session replay, ethically. Is it okay to do? Do you think visitors should be informed about it? Would you use a website differently if there was a big red banner saying your cursor movements and viewport are being livestreamed to HQ?
Genuinely curious, because it's not something I think most people are aware of when they browse the web.
vasco 27 minutes ago [-]
I've worked on such a product for ~8 years (one of the most widely adopted ones) and find that most of these sessions are used in practice to fix bugs or improve UX by such a vast majority that even if there's some voyeuristic aspect I don't think it's super important to care about. The product I worked on made no effort to try and bypass ad blockers or browser settings like DNT, and a lot of sophisticated users wouldn't be recorded anyway due to those settings. We didn't even do any fingerprinting and auto-masked all fields. Overall it was less privacy invading than walking in a shop while being recorded by a security camera in my opinion. I don't work in this field anymore but would do it again.
chrisandchris 2 hours ago [-]
How does this compare to e.g. OpenReplay, which looks quite similar?
OpenReplay is much more mature and full-featured, RePlaya is just the core session capture, listing, and replay functionality. OpenReplay has more dependencies, so self-hosting means running a full stack: Postgres, ClickHouse, Redis, and its backend services. RePlaya is one stateless Node process plus S2 (or self-hosted s2-lite).
rgbrgb 4 hours ago [-]
very cool. session replays are so crucial for understanding new features qualitatively, especially before you have enough users for useful quantitative metrics.
ballpark, how much does this cost to run?
shikhar 2 hours ago [-]
Thanks! And agreed, session replays can be really useful to understand user behaviour such as product edge cases.
On cost, it's running the collector Node app (I'd expect a few $ per month at low volume), and the S2 stream backend.
If you use the S2 cloud service, cost is basically just the rrweb bytes. The rates are $0.075/GiB to write, $0.05/GiB-month to store, $0.10/GiB to read back over the internet. See s2.dev/pricing.md for an agent-friendly summary.
Assuming a typical few-minute session is ~1 MiB of events, ingesting it, storing it a month, and replaying it a couple of times (unlikely!):
Genuinely curious, because it's not something I think most people are aware of when they browse the web.
[1] https://www.openreplay.com
ballpark, how much does this cost to run?
On cost, it's running the collector Node app (I'd expect a few $ per month at low volume), and the S2 stream backend.
If you use the S2 cloud service, cost is basically just the rrweb bytes. The rates are $0.075/GiB to write, $0.05/GiB-month to store, $0.10/GiB to read back over the internet. See s2.dev/pricing.md for an agent-friendly summary.
Assuming a typical few-minute session is ~1 MiB of events, ingesting it, storing it a month, and replaying it a couple of times (unlikely!):
1k sessions/mo ≈ $0.35
10k sessions/mo ≈ $3.50
100k sessions/mo ≈ $35