PAGE_SLEEP_MIN being ignored?

[daviderickson]

Description

Hi running on Windows 10, my .env has configured PAGE_SLEEP_MIN (and PAGE_BACKOFF_MIN) both to 60000, so I'd expect a 60 second minimum delay between queries to the same store. However from the log printout I'm getting:

[12:18:06 PM] info :: ✖ [nvidia] [nvidia (3080)] founders edition :: OUT OF STOCK
[12:18:10 PM] info :: ✖ [nvidia] [nvidia (3080)] founders edition :: OUT OF STOCK
[12:19:13 PM] info :: ✖ [nvidia] [nvidia (3080)] founders edition :: OUT OF STOCK
[12:19:17 PM] info :: ✖ [nvidia] [nvidia (3080)] founders edition :: OUT OF STOCK

Which looks to me like the min value isn't being respected by the backend.

Thanks!

[xninjax]

Hi @daviderickson, just wondering if your running the latest nvidia-snatcher? if you haven't done an update in a while, it may be worth doing so, as this issue should have been resolved in #465 (comment).

Alternately, you can try setting PAGE_SLEEP_MAX to 60000 as well, as this would be a workaround fix (if you haven't pulled a recent update/ or don't want to yet).

Let us know if either of these resolve the issue for you 👍

[jef]

Going to close and we can reopen if it ends up being a regression. Thanks @xninjax.

[daviderickson]

@xninjax @jef Yeah I pulled yesterday, I'm on f24f565 same problem.

[andrewmackrodt]

Looks like intended behaviour, the 3080 FE has 2 entries using different pages IIRC, same as the 3090. I was also confused by this behaviour when I first used the project, like @daviderickson I expected PAGE_SLEEP_MIN to be a 60 second minimum delay between queries to the same store. However, it acts as a delay between batches of queries to the same store.

[jef]

However, it acts as a delay between batches of queries to the same store.

That's correct. That being said, I was curious to make these requests to all the links parallel, since we aren't setting intervals between them. But not sure of the side effects.

[xninjax]

Agreed the multiple hits can appear broken/confusing at first, but indeed the script is confirmed only checking one set of links per series per batch (i.e every 60sec here). 2 links per series on nvidia sore ✔️

• logger.info delay added for debug purposes

EDIT:

Perhaps the PAGE_SLEEP_MIN/MAX definition could be improved to better reflect its current intended usage?

PAGE_SLEEP_MIN | Minimum sleep time between queries of the same store | In milliseconds, default: 5000
PAGE_SLEEP_MAX | Maximum sleep time between queries of the same store | In milliseconds, default: 10000

Change to:

PAGE_SLEEP_MIN | Minimum sleep time between queries of the same product page | In milliseconds, default: 5000
PAGE_SLEEP_MAX | Maximum sleep time between queries of the same product page | In milliseconds, default: 10000