POST requests suddenly began timing out

I’m running Trellis 0.9.8 / Bedrock / Sage 8.5 on a Digital Ocean Ubuntu 16.04 1cpu/512mb droplet. I’m posting here because I’m hoping one of the fine minds here might have some insight on my issue and DO is being unresponsive.

The website sends POST requests to two external APIs (using wp_remote_post I believe). Starting yesterday, all POST requests have started timing out. All other functionality seems to be unaffected. I’m not sure where to start on debugging this exactly. Would it be a limit issue with the DO plan? I haven’t made any changes in the 3 months the site has been running other than updating WP core and plugins. None of the updates were made in the last week that would suggest a direct cause.

Does restarting the PHP service fix the problem?

No, I did try that, but no change.

Here’s an error message from the timeout with the Wordfence API for example:

wp_remote_post() test to noc1.wordfence.com failed! Response was: cURL error 28: Resolving timed out after 10523 milliseconds

“Resolving” should mean a DNS issue. We can’t really help much beyond that. It’s likely not an issue with Trellis itself.

You can try dig noc1.wordfence.com from the command line and debug from there.

Yeah I was assuming it was probably not, but given that DO seems widely used by everyone on this forum I was hoping someone might have a similar issue. Always appreciate advice from all you guys on this site.

For anyone who might be experiencing something similar, looks like it is an issue with Google’s DNS servers resolving UDP queries from a specific DO IP block. Here is their published fix.

1 Like