Opened 4 years ago

Last modified 13 days ago

#17449 new defect

Test bridges we ship in Tor Browser regularly

Reported by: gk Owned by: boklm
Priority: Medium Milestone:
Component: Applications/Quality Assurance and Testing Version:
Severity: Normal Keywords: tbb-usability, tbb-bridges
Cc: phw Actual Points:
Parent ID: Points:
Reviewer: Sponsor: Sponsor30-can

Description (last modified by gk)

We should try to detect broken bridges we ship in Tor Browser early to minimize users trying them and to avoid overloading the remaining ones considerably/over a long period of time.

Not sure what "regularly" should mean in this context but I guess just testing them during our normal QA might not be enough as they could be down/unreachable temporarily.

We might set up something different maybe, even with a system that sends an email to the bridge operator (Cc'ing tor-assistants) after the bridge is repeatedly not usable.

Child Tickets

Change History (3)

comment:1 Changed 4 years ago by gk

Description: modified (diff)

comment:2 Changed 13 days ago by pili

Cc: phw added
Sponsor: Sponsor30-can

comment:3 Changed 13 days ago by phw

Keywords: tbb-bridges added

Thanks to gman999, we now have a sysmon instance that continuously tests our default bridges. Our sysmon config is available in this gitlab repository. Here's how it works:

Sysmon runs checks every five minutes, and updates DNS records every ten minutes. If a check fails twice (e.g., a service is offline for more than five minutes), we get an alert. Currently, cohosh and phw receive alerts. Our sysmon instance currently does not have IPv6 capabilities.

Whenever one of our default bridges fail, cohosh and I get an email. This has happened a few times already. Do we want anything else?

Note: See TracTickets for help on using tickets.