Slow running tests and Lighthouse failures

[cpk]

Did something change in the agent behavior maybe, around 2/8 noon pst? My private instance agents started to queue badly with no change in test load. I've added 50% more agents and they still can't keep up. Simultaneously Lighthouse scores stopped being calculated. The test log gives this message and shows the job being killed

Sorry, Lighthouse had some issues gathering your report on WebPageTest. Please try again or try using Lighthouse through another way

I can repro this on the public instance

https://www.webpagetest.org/lighthouse.p...e2a7&run=1

[sertovs]

Are you sure you have chosen a test agent that supports running Lighthouse? Check out this post. We have a private instance where we have a linux test agent (one of the linux options here - only they can run Lighthouse). It runs Lighthouse successfully if you choose to use mobile emulation as well.

[cpk]

Yes. I am using the latest Linux AMIs. They have been getting lighthouse scores with 100% success until 2/8 at noon and now they are 100% failures. I also had a repro on the public instance.

[cpk]

looks like the lighthouse failures are all due to a busted lighthouse 2.9.0 2.9.1 was just released a few minutes ago.