Creating a Snapshot API is down

[aylusltd]

https://archive.readme.io/docs/creating-a-snapshot

Example returns 502 bad gateway error. Has been down for several days.

[mekarpeles]

Hi @aylusltd thank you for the heads up. I should have time to debug this weekend (we've had a lot going on)

In the meantime you may wish to use the official Wayback SPN 2.0 (save page now)
https://help.archive.org/hc/en-us/articles/360001513491-Save-Pages-in-the-Wayback-Machine#:~:text=Install%20the%20Wayback%20Machine%20Chrome,give%20you%20a%20permanent%20URL.

[aylusltd]

@mekarpeles , I can't actually use the extension. We use the API form a lambda to get back the permanent url and queue it up for human annotation.

I guess we could try to automate that, but seems like it'd be easier for us just to read the extension code and match its API invocation.

[mekarpeles]

Should be back up, I think :)
The annotations for this labs is fairly experimental (and not maintained by IA). If anyone is relying on this for mission critical purposes (i.e. you'd be upset if annotations db disappeared) I suggest running this service locally or submitting a PR for a simple docker setup.

This code doesn't require any privileged setup to run and installation is pretty simple:
https://github.com/ArchiveLabs/pragma.archivelab.org#installation

[mekarpeles]

#21 (comment)

[jerclarke]

I just tried the example from https://archive.readme.io/docs/creating-a-snapshot and got a 502 error.

Maybe it's just the example that's the problem, but figured I'd mention it.

What I really want is the old system where a call to http://web.archive.org/save/http://google.com would create a cache of the URL, and return a Content-Location header with the /web/* path of the archive, but I think I'm probably in the wrong place for that. That "trick" stopped working July 10 (bug report on the AmberLink project).

[mekarpeles]

If https://pragma.archivelab.org ever throws a 502, that means the service is down. Maybe it's getting DDOS'd? I just restarted the service and it looks like it's currently working.

Sincere apologies for playing close/open tag on this issue! @jerclarke or @aylusltd -- please feel free to close if this resolves your issue :) I'll defer to you two and be on standby if I can help further.

[Towito]

Hi, I've been trying to use this API and it appears to be down. The GET snapshot API appears to be fine, but POST requests do not appear to be working. I'm hoping to go through multiple web pages on a somewhat regular basis, so the SPN function is not particularly useful to me. Additionally, the size of webpages is far too small to justify an Archive-IT solution. Has this service been deprecated or something?

[mackuba]

Yeah, not working for me either…

[SemjonWilke]

Still 502

[mekarpeles]

Here's the deal --

pragma.archivelab.org was intended to be a system for saving Wayback snapshots with annotations attached. Very few snapshots have had what I'd consider descriptive, meaningful, or specific annotations attached. Leading me to believe most people just want the use the Wayback API (of which there is one).

The problem here is, I now have a database of 17,000,000 urls that y'all have archived and postgres can't keep up.

So, the course of action I'm planning is that this code be changed to just proxy using the Wayback API and not write to the database at all, unless an annotation is added.

Given the current performance characteristics and use -- which I had not intended / planned on, given that I wrote this while I was a volunteer not even working for the Internet Archive -- I have to now consider what to do with all of this data...

[mekarpeles]

Ok, I added this capability:
https://github.com/ArchiveLabs/pragma.archivelab.org/blob/master/README.md#performing-a-capture

curl -X POST -H "Content-Type: application/json" -d '{"url": "https://google.com"}}' https://pragma.archivelab.org/capture

[Pandaklez]

It still throws 502 error :(

[ronthepennyhoarder]

Getting a 502 error as well