James Titcumb

Note: `roave-backwards-compatibility-check.sh` script can probably be dropped now too.

@andeewoodward hey! Great suggestion. Out the box, there's nothing automatic to only enable on %age of requests, but you could simulate that by changing the monitoring enabled environment variable `SCOUT_MONITOR`...

Note: this is definitely an older version of `scout-apm-php` since the `scout-apm-laravel` code exists. Not sure which version is being used though.

Further info; changing the env var may be the cause if this, depending on how Heroku behaves when this happens; perhaps it kills the running process, but leaves the `/tmp`...

Hmm, did some investigation this afternoon, I'm not sure changing the env var is the cause here; I tried this in Heroku and it did fully restart the node: ```...

Hmm, still unable to reproduce this. On my test Heroku instance I added an HTTP endpoint that kills the running `core-agent` (on the web node, not an ephemeral CLI node!),...

Now at least we are not "crashing" with OOM errors, we can investigate the underlying cause. First step is to try to reproduce the issue. @jrothrock mentioned customer is not...

@pesseyjulien The error occurs because the `scout-apm-symfony-bundle` library cannot connect to something called `core-agent`. The `core-agent`, with the default configuration, is automatically downloaded and launched on the first request to...

@pesseyjulien would you be able to provide that configuration that should be displayed in the log output please, as that should help us determine if the messages are getting sent...

> I'm sorry but it seems there is no way to see this kind of log on Heroku :/ > > But I revert the SCOUT_LOG_LEVEL to DEBUG and got...