Bridgestrap tests not syncing to colchicifolium

Hey, I have noticed the bridgestrap tests are not being synced to colchicifolium. Was this intended?

Cheers

marked this issue as related to tpo/network-health/metrics/collector#40042 (closed)

assigned to @meskio

added Doing For Network Health Team labels

https://bridges.torproject.org/bridgestrap-collector and /home/bridgestrap/logs/collector-metrics.log in polyanthum are empty.

The log rotated files are also empty, the oldest one is from April 18. And in collector seems to be that the last fetched file is from April 16.

The last deployment of bridgestrap was March 20th, so it doesn't match with that date. But the uptime says 8 days, so it does matches April 16. Looks like bridgestrap has stopped providing collector metrics since the last reboot of polyanthum. Weird that we didn't see this problem after deploying it, could it be that I copied the binary into the server and didn't restart the service?

mmm, restarting the service has solved the problem. I'm very confused now.

Looks like the problem is back and bridgestrap is reporting a blanc collector file.

I believe the problem is coming from me forgetting to add copytruncate into the [logrotate.conf]( bridgestrap-admin@dc4eec33), and logrotate is now emptying the collector file. We didn't notice this change before the reboot of the server as the logrotate configuration was not applied until then.

I'm checking with TPA how to apply this configuration, not sure I have rights to do so.

bridgestrap is still publishing metrics after the change. I think the problem is solved. I'll close this issue.

closed

Hey @meskio Unfortunately it's not fixed:

Actually, it looks like bridgestrap is reporting that bridge to metrics correctly.

$ wget https://collector.torproject.org/recent/bridgestrap/2024-05-03-00-40-04-bridgestrap-stats
[...]
$ grep -i f6eb9fcf 2024-05-03-00-40-04-bridgestrap-stats 
bridgestrap-test true F6EB9FCFA5FE54CBE847D993D55EE53725D136C2

So, bridgestrap is doing its job, and it is sending the data to collector properly.

You might reasonably reply "ok fine but my bridge is still marked as down, something's wrong" -- and you are right. The next step is to wonder why the relay-search page chooses to label you as Down.

Well, I apologise, I mean I don't have deep knowledge on how the backend works and how each system is connected with the others, I just wanted to notify you on the (related ?) issue.

Relay search doesn't show me always down though, randomly it shows me up, with the correct uptime.

Then, suddenly, it shows me down with a random downtime.

@Deny thank you for reporting it, is good that you reported it here, we can move it to the right place when we are sure on the problem.

@hiro can you look if the problem is in your side?

Yes I think I will open an issue on onionoo.

tpo/network-health/metrics/onionoo#40046

Great, I'll close this issue then. Let me know if you need any help from me there.

reopened

added Next label and removed Doing label

mentioned in issue tpo/network-health/metrics/onionoo#40046

marked this issue as related to tpo/network-health/metrics/onionoo#40046

closed