Fork and alt host interest?

[berezovskyi]

Hello,

https://sparqles.ai.wu.ac.at/ which to my best knowledge was the main endpoint of SPARQLes, is down for quite a long time now. Additionally, the software is quite old and needs at least a refresh of dependencies and a dockerization. I have two questions:

1. Does anyone have interest in helping maintain a fork of this repo? Main aim would be to do low-key dependency updates to catch up with security updates and new formats/protocols.
2. Does anyone know of an existing instance of SPARQLes running elsewhere? I am interested from the point of asking for MongoDB backups such as to retain the uptime history as much as possible.

Cheers, Andrew

[berezovskyi]

So far, I have dockerized the whole setup using Docker Compose and got it to run again (I forgot it since I ran it 5 years ago). Almost 600 endpoints loaded from datahub.io and just short of 20% of them are alive.

I also registered the sparqles organization on Github to transfer the repo. For now, I am planning to transfer my repo (berezovkyi/sparqles) over there unless @pyvandenbussche or others suggest otherwise. My plan is to mitigate biggest security holes by updating some libraries to the supported versions and deploy the server on one of the cheapest VMs on Hetzner (2vCPU/2GB RAM/40GB disk at ~€5/mo) and see how it goes (I am also thinking of disabling performance checks to avoid hammering servers with unnecessary load; we may also want to split robots.txt agents for general uptime checks and the rest of extended checks). Some immediately relevant issues on this repo are:

• Better server ident, see #26 #48 #62. On top of what's written there, most triplestores implement at least some SPARQL extensions that can be sniffed, e.g. Jena's ARQ filters and Blazegraph has FTS extensions.
• Some process for adding/updating endpoints, esp. given that databus.io seems to be dormant? See #42 #50, for example. I am thinking of a w3id-like approach, where people can make PRs to a text file and it's then loaded by the server. This should also make SPARQLes useful for intranet deployments where many endpoints may need to be monitored.
• Is there a need for notifications? The endpoints that I monitor myself are now simple HTTP GET checks every 5 minutes with an ASK query sparql?query=ASK%20{?s%20?p%20?o} set up in UptimeRobot or Uptime Kuma:

However, my knowledge of Node and Mongo is quite limited. Thus I am asking here if someone would be interested to help or point me to some other effort.

[AxelPolleres]

Hello,

https://sparqles.ai.wu.ac.at/ which to my best knowledge was the main endpoint of SPARQLes, is down for quite a long time now. Additionally, the software is quite old and needs at least a refresh of dependencies and a dockerization. I have two questions:

1. Does anyone have interest in helping maintain a fork of this repo? Main aim would be to do low-key dependency updates to catch up with security updates and new formats/protocols.

2. Does anyone know of an existing instance of SPARQLes running elsewhere? I am interested from the point of asking for MongoDB backups such as to retain the uptime history as much as possible.

Cheers, Andrew

Hi Andrew,

it's up again now and was down mainly for log4J issues (fixed now) ... we'd be happy to join forces if you want to contribute. I think though, it's not a good idea to run it on two different sites/servers ... and we're still happy to "sponsor" the running server from our institute. Let us know what you think, we'd be happy to have a call and sort out the best way to collaborate/join forces!

Axel