-
Notifications
You must be signed in to change notification settings - Fork 142
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
webconnectivity: handle down websites such as doh.centraleu.pi-dns.com #2299
Comments
So, we're fixing this for v0.5 inside ooni/probe-cli#957. It remains to be seen whether we also want to apply a similar fix to the v0.4 implementation, or whether we prefer focusing on v0.5. I'll chat with @hellais to get a second opinion about this topic. I'd ideally like to get more experience with v0.5 and then try to bring some changes back to v0.4 in a batch, because it might be more efficient to do that. |
#957) This diff introduces a special rule to avoid emitting null, null when all the connects failed in both the probe and the TH. While there, recognize that the subset of null, null we're hunting actually deals with websites that are down, so change the internal naming to reflect that and make the code easier to read/understand. See ooni/probe#2299
…website down (#957) This diff introduces a special rule to avoid emitting null, null when all the connects failed in both the probe and the TH. While there, recognize that the subset of null, null we're hunting actually deals with websites that are down, so change the internal naming to reflect that and make the code easier to read/understand. See ooni/probe#2299
I just added the fixed-by-webconnectivity-lte flag such that we can flag all the issues (such as this one), that are open but will automatically be fixed when we switch over to the new implementation of Web Connectivity. I am now going to tag similar issues. Because the development of the new version of Web Connectivity took a long time, we need a way to distinguish between issues that have been fixed in the next version of Web Connectivity and issues that have not been fixed yet. |
This diff corrects an embarassing bug in the logic we were using for expected TCP connect and TLS handshakes failures. We spotted this issue with the new testcase we added called websiteDownTCPConnect, which is also included. Additionally, we needed to fix some tests because the logic bugs were preventing us from seeing some events. Part of ooni/probe#2652. Closes ooni/probe#2299.
This diff corrects an embarrassing bugs in the logic we were using for expected TCP and TLS failures. We spotted this issue with the new testcase we added called websiteDownTCPConnect, which is also included. Additionally, we updated some tests because of the embarrassing bugs we fixed. Additionally, we fixed some comments that were outdated. Part of ooni/probe#2652. Closes ooni/probe#2299.
This diff corrects an embarrassing bugs in the logic we were using for expected TCP and TLS failures. We spotted this issue with the new testcase we added called websiteDownTCPConnect, which is also included. Additionally, we updated some tests because of the embarrassing bugs we fixed. Additionally, we fixed some comments that were outdated. Part of ooni/probe#2652. Closes ooni/probe#2299.
Let us start our journey with the obligatory MAT search:
So, judging from the MAT search, this website presents anomalies. But, is that really the case? Could it be that there is a bug in Web Connectivity v0.4.x that causes the website to erroneously being flagged as anomaly?
Judging from this v0.4.1 measurement maybe we're misflagging the results here. The connections indeed all fail for both the probe and the TH, so this should be flagged as "website down", at least as far as that measurement I showed was concerned. But also this measurement shows similar issues. And also this one. So maybe I have a point here 🥶.
The approach taken by v0.5 LTE for this kind of situations is a bit more correct. In this measurement, it says "look, mate, I don't know", and sets .Blocking to null and .Failure to null, thus making it an errored measurement. However, the best approach in this case would be to recognize that the website is just down and flag is as such like we did for DNS a few commits ago in ooni/probe-cli@b10eea4.
The text was updated successfully, but these errors were encountered: