[BUG] ARC HPC NAT IP Address Blocked: ConnectionResetError when Accessing EnrichR API

[xuxu-wei]

Hi EnrichR team:

I'm a researcher at University of Calgary. We are encountering connection errors when trying to access the Enrichr API from the University of Calgary’s ARC HPC environment. The specific error message we receive is: ConnectionResetError: ('Connection aborted.', ConnectionResetError(104, 'Connection reset by peer'))

This issue seems to be specific to the ARC HPC environment, as the same API calls work fine on local machines and other environments. Upon investigation by the ARC administrators, it appears that the ARC NAT IP address (136.159.79.11) may be blocked by Enrichr’s server.

Could you please check if the ARC HPC NAT IP (136.159.79.11) has been blocked on your server? We would appreciate any assistance in resolving this issue.

• Recent Attempts: Both I and my labmate were able to successfully use the Enrichr API from ARC in the past, with the most recent successful attempts being on 2024-09-11. This issue seems to have arisen recently.

• Environment Details: ARC NAT IP: 136.159.79.11 Python: 3.9.13 requests: 2.28.1 urllib3: 1.26.11 certifi: 2022.9.14 OpenSSL: 1.1.1q

• Network Connectivity: Used curl and wget to test connectivity to the Enrichr server. The output is attached Network_Connectivity.log.

• Tested SSL Configuration in Python: Attempted to disable SSL verification in the Python requests call (verify=False), but the error persisted. Tried forcing TLS 1.2 usage by customizing the SSL context, but this did not resolve the issue. Attempted to send request using urllib rather than requests. The outputs are attached.

[AviMaayan]

Dear @xuxu-wei we had to block you because you were submitting too many requests per second...

[xuxu-wei]

Dear @xuxu-wei we had to block you because you were submitting too many requests per second...

Dear @AviMaayan

Thank you for your prompt response and for clarifying the reason behind the block.

Could you please advise on the recommended number of requests and the appropriate frequency for interacting with the Enrichr API? I will make sure to inform everyone in our lab to ensure that we adhere to the guidelines and avoid placing too much load on your servers.

Additionally, would it be possible to have our access to the API reinstated? The ability to use the Enrichr API is critical for our ongoing research, and we will take the necessary precautions to avoid overwhelming the server in the future.

We greatly appreciate your support and understanding.

[AviMaayan]

Great. Please limit the requests for one per second.

[xuxu-wei]

Great. Please limit the requests for one per second.

Thanks! I will tell all members of our lab to limit requests to one per second moving forward. We truly appreciate your support.

[u8sand]

Just writing to let you know that this ip address has been unblocked for now. Thank you for exercising restraint when accessing our services which are free resources depended on by many other researchers.

We do have another service for performing enrichment analysis over API exclusively: https://maayanlab.github.io/smartAPIs/speedrichr/

Which should be able to handle a higher load, still, please avoid making too many requests per second.

[xuxu-wei]

Just writing to let you know that this ip address has been unblocked for now. Thank you for exercising restraint when accessing our services which are free resources depended on by many other researchers.

We do have another service for performing enrichment analysis over API exclusively: https://maayanlab.github.io/smartAPIs/speedrichr/

Which should be able to handle a higher load, still, please avoid making too many requests per second.

Dear @u8sand, Thanks so much for unblocking our IP and for sharing the Speedrichr API link!

I have a quick question regarding Speedrichr: on the EnrichR help page, I see that it can be used for enrichment analysis with a background gene set, but I’m not sure if adding a background gene set is always required. If I don’t include a background gene set, will Speedrichr behave the same as https://maayanlab.cloud/Enrichr/?

In other words, would I just need to replace https://maayanlab.cloud/Enrichr/ with https://maayanlab.cloud/speedrichr/ in my requests, and everything else will work the same?

Thanks again!

[u8sand]

The primary difference is that speedrichr does not persist results indefinitely like Enrichr does. It should work as you described without needing to provide a background. There is a separate endpoint when you want to enrich with a supplied background.

[xuxu-wei]

Perfect! Thank you for the clarification! I will try to proceed with Speedrichr as suggested.