Christoph Kutzinski

Would be really really nice to have some progress here. I've just been bitten by this bug and it's really hard to catch as the numbers jq spits out looks...

Any update on this issue? Would be nice if we could use the latest release with Docker @connorjclark ?

Can confirm that server starts and is able to read old data volume. Will now check if publishing with the new client works

lhci-client doesn't look so good: ``` [2022-03-24T15:45:14.823Z] Running Lighthouse 5 time(s) on ... [2022-03-24T15:48:24.975Z] Run #1...failed! [2022-03-24T15:48:24.975Z] Error: Lighthouse failed with exit code 1 [2022-03-24T15:48:24.975Z] at ChildProcess. (/usr/local/lib/node_modules/@lhci/cli/src/collect/node-runner.js:120:21) [2022-03-24T15:48:24.975Z] at...

Didn't change anything except the image. With the old lhci-client image it was and still is (against the new server) working.

Seems to be related to https://github.com/GoogleChrome/lighthouse/issues/6512#issuecomment-1029150960 I've added the custom chrome option `--disable-dev-shm-usage` as suggested and now it seems to run stable

> In your case, which database option do you use? I just started it with the standard options given in the readme: `docker container run --publish 80:9001 --mount='type=volume,src=lhci-data,target=/data' --name lhci-server...

> Docs.... yeah At least mentioning it at all in the docs about supported store technologies would be nice ;-)

FYI: I've created a PR for configuring the key prefix in consul: https://github.com/ff4j/ff4j/pull/525

@daniel-beck good point about overlapping directories. I guess the behaviour would be unpredictable in worst case. I'm already checking for exact duplicates and issue a warning. Will try to do...