Some of our requests trigger large clean ups in our Redis cache causing the creation of a large amount of child spans. This in turn creates thousands of stderr log lines making the stderr log unusable. We can't use ignoreUrls because many requests to the same URL are ok and we want to trace those. Also we can't know before hand how many redis del commands there will be, it could be tens of thousands. How can we avoid this? Ideally we would just want to limit the number of child spans and not log anything.

The log lines looks like this:

@google-cloud/trace-agent ERROR TraceApi#createChildSpan: [[email protected]] Creating phantom child span [redis-del] because the trace with root span [/some/url] has reached a limit of 3200 spans. This is likely a memory leak.