Jonathan Gray
Jonathan Gray
Well, any scenario where docker swarm isn’t the deploy method really. Or where it is, and you want something more robust for credential rotation for example.
I’ve dropped codecov an email, because it honestly looks to me like this is something wrong at their end.
I guess this is really two issues. The first, is that a flowmachine can fail and write a bad cache record as a result of hitting the recursion limit when...
One can 'fix' this by increasing the recursion limit of course, but that's maybe only delaying the inevitable (I triggered this by nesting a few hundred queries).
Let's start with adding a benchmark for this?
In conjunction with reduced tables work.
This should make use of the weight field in the geo_bridge table under the new schema
Extending from this idea.. I think we might consider splitting flowmachine into three - - flowmachine, which contains only the necessaries for running queries against flowdb - flowmachine-queries (or Flowmachine-features),...
I guess ideally it'd start on a random free port and tell you which one.