HPR/hpr_hub
5
0
Fork 2
Code Issues 45 Pull Requests Packages Projects Releases Wiki Activity

Server Bandwidth Very High - And High Costs #88

New Issue
Open
opened 2025-02-24 04:04:08 +00:00 by jknapp · 11 comments
jknapp commented 2025-02-24 04:04:08 +00:00
Owner
Copy Link

Hey folks,

I have been moving systems out of AWS to try and reduce the ballooning costs, and I have now reached the point I am auditing HPR's usage.

For the last few months at least, HPR has been using between 100 to 200 USD in bandwidth alone. I have set up servers in another Data Center that does not charge for bandwidth, which has reduced my costs significantly for other systems.

The good news means I can assign more CPU and RAM for the instance than what is on the current system.
The bad news, we have to re-setup the server in the new DC.

I would like to start working on this soon. As the total costs for the HPR server is closing in on about $300 USD per month, and donations have been pretty dry. Donations being dry is fine, I understand it is a tough time for all.

If you have requests or questions, please let me know.

P.S. if you want to donate to me for HPR, you can do so here.

Hey folks, I have been moving systems out of AWS to try and reduce the ballooning costs, and I have now reached the point I am auditing HPR's usage. For the last few months at least, HPR has been using between 100 to 200 USD in bandwidth alone. I have set up servers in another Data Center that does not charge for bandwidth, which has reduced my costs significantly for other systems. The good news means I can assign more CPU and RAM for the instance than what is on the current system. The bad news, we have to re-setup the server in the new DC. I would like to start working on this soon. As the total costs for the HPR server is closing in on about $300 USD per month, and donations have been pretty dry. Donations being dry is fine, I understand it is a tough time for all. If you have requests or questions, please let me know. P.S. if you want to donate to me for HPR, you can do so [here](https://liberapay.com/GoTakeAKnapp).
rho_n commented 2025-02-24 16:41:01 +00:00
Owner
Copy Link

Hi Josh,

Can you give us any breakdown on the usage? Is it just general downloads of our audio? Is there any particular ip/region/etc that is using more than others? It would be helpful in determining if it is something we are inadvertently doing that is causing high-bandwidth usage.

Thanks for all you do for the HPR community!

Hi Josh, Can you give us any breakdown on the usage? Is it just general downloads of our audio? Is there any particular ip/region/etc that is using more than others? It would be helpful in determining if it is something we are inadvertently doing that is causing high-bandwidth usage. Thanks for all you do for the HPR community!
jknapp commented 2025-02-26 13:44:18 +00:00
Author
Owner
Copy Link

Hi @rho_n

Unfortunately AWS doesn't really give me too much detail as to what is consuming the bandwidth, just that it has been used and they want their pound of flesh.

I did log into the server and look to see if there was anything obvious that would be consuming so much bandwidth, and i did not see anything I could be certain about causing the usage.

I know HPR switched to using Archive.org for streaming the audio. How are the files encoded and sent to Archive.org?

Hi @rho_n Unfortunately AWS doesn't really give me too much detail as to what is consuming the bandwidth, just that it has been used and they want their pound of flesh. I did log into the server and look to see if there was anything obvious that would be consuming so much bandwidth, and i did not see anything I could be certain about causing the usage. I know HPR switched to using Archive.org for streaming the audio. How are the files encoded and sent to Archive.org?
ken_fallon commented 2025-02-28 17:48:38 +00:00
Owner
Copy Link

Thanks @jknapp for looking into it. All processing is done offline, the files are uploaded to hpr and I pull them from there and send them from my PC. I'm going to pull the logs and analyze them offline to see what is going on. I did notice a lot of AI bot scraping going on but let's wait and see.

In any event we need to move as there is no point paying more than we need to.

What is the plan ? Is https://repo.anhonesthost.net also moving ?

Thanks @jknapp for looking into it. All processing is done offline, the files are uploaded to hpr and I pull them from there and send them from my PC. I'm going to pull the logs and analyze them offline to see what is going on. I did notice a lot of AI bot scraping going on but let's wait and see. In any event we need to move as there is no point paying more than we need to. What is the plan ? Is https://repo.anhonesthost.net also moving ?
jknapp commented 2025-03-04 16:29:39 +00:00
Author
Owner
Copy Link

Repo is already moved over, I did that one a while ago as it was a pretty easy transition. I will get a new server stood up and share with you the login details for the new machine.

I do have a control panel I am working on developing that works with containers, if you guys do not want to stand up the config for the web server again.

Repo is already moved over, I did that one a while ago as it was a pretty easy transition. I will get a new server stood up and share with you the login details for the new machine. I do have a control panel I am working on developing that works with containers, if you guys do not want to stand up the config for the web server again.
norrist commented 2025-03-05 12:49:07 +00:00
Member
Copy Link

I looked at the apache logs with goaccess - https://goaccess.io/man
It shows 642.73 GiB served from apache 02/Feb/2025 — 05/Mar/2025.
Does this account for most of the bandwidth used, or do we need to look at another service?

I will attach the HTML output from goaccess. The IP addresses have been anonymized, I can look further into a specific IP address if needed.

goaccess access_log* -o report.html --log-format=COMBINED --all-static-files --anonymize-ip --anonymize-level=3

I looked at the apache logs with goaccess - https://goaccess.io/man It shows 642.73 GiB served from apache 02/Feb/2025 — 05/Mar/2025. Does this account for most of the bandwidth used, or do we need to look at another service? I will attach the HTML output from goaccess. The IP addresses have been anonymized, I can look further into a specific IP address if needed. `goaccess access_log* -o report.html --log-format=COMBINED --all-static-files --anonymize-ip --anonymize-level=3 `
report.zip
254 KiB
ken_fallon commented 2025-03-06 11:04:27 +00:00
Owner
Copy Link

Great work. We need to look at the full feeds to see if we can reduce them but other than that it seems to be business as usual.

Great work. We need to look at the full feeds to see if we can reduce them but other than that it seems to be business as usual.
jknapp commented 2025-03-06 21:23:24 +00:00
Author
Owner
Copy Link

That is awesome, I did not know of that tool. The place I am moving stuff to does not bill for bandwidth, so that is a bonus.

That is awesome, I did not know of that tool. The place I am moving stuff to does not bill for bandwidth, so that is a bonus.
jknapp commented 2025-03-06 22:44:08 +00:00
Author
Owner
Copy Link

Also, does anyone have a preference to the flavor of linux that is used? I can do Debian 12, Almalinux 9, or Ubuntu 24.04.2 LTS. It can be bare or I can toss on the docker based control panel I am working on.

Also, does anyone have a preference to the flavor of linux that is used? I can do Debian 12, Almalinux 9, or Ubuntu 24.04.2 LTS. It can be bare or I can toss on the docker based control panel I am working on.
ken_fallon commented 2025-03-11 17:46:45 +00:00
Owner
Copy Link

I'd prefer not to have Ubuntu as they tend to be a bit weird in their choose of stuff. A control panel thing would be great but access to the server would also be great.

The biggest pain the last time was the email.

I'd prefer not to have Ubuntu as they tend to be a bit weird in their choose of stuff. A control panel thing would be great but access to the server would also be great. The biggest pain the last time was the email.
jknapp commented 2025-03-12 12:41:18 +00:00
Author
Owner
Copy Link

I'll set up a server with Alma9 then. How is mail being done now? It is just the mailing list being the problem right?

I'll set up a server with Alma9 then. How is mail being done now? It is just the mailing list being the problem right?
ken_fallon commented 2025-03-12 13:12:22 +00:00
Owner
Copy Link

The mailing list is now on a different domain lists.hackerpublicradio.org and is a managed service by mailman.net

The mailing list is now on a different domain lists.hackerpublicradio.org and is a managed service by mailman.net
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
No items
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
davmo jknapp ken_fallon norrist rho_n
No Assignees
4 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: HPR/hpr_hub#88
No description provided.
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?