Database structure changes can cause the cluster to fail to serve.

[wesbillman]

If we run ftl dev . in ftl-examples/online-boutique everything starts up correctly:

ftl dev online-boutique/backend --recreate
info: Starting FTL with 1 controller(s)
info:currency: Building module

# ... building and deploying...

info:controller0: Deployed cart-8c85f8485b
info:controller0: Deployed checkout-90b6ae9e13
info:controller0: Deployed recommendation-880c13394c
info:controller0: Deployed shipping-ed10711eb0
info:controller0: Deployed productcatalog-ff266169ab
info: All modules deployed, watching for changes...

If I then go start the server from a new folder (like ftl) and run ftl serve without recreating anything, I get the following warnings and errors

ftl serve

info: Starting FTL with 1 controller(s)
info:controller0: Web console available at: http://localhost:8892
warn:reconcileRunners: failed to get deployments needing reconciliation: can't scan into dest[4]: 14:3: ingress verb getQuote: couldn't resolve response body type currency.Money
warn:reconcileRunners: failed to get deployments needing reconciliation: can't scan into dest[4]: 14:3: ingress verb getQuote: couldn't resolve response body type currency.Money
warn:reconcileRunners: failed to get deployments needing reconciliation: can't scan into dest[4]: 14:3: ingress verb getQuote: couldn't resolve response body type currency.Money
error:controller0: Failed to increase deployment replicas: failed to claim runners for checkout-90b6ae9e13: no idle runners found matching labels {"languages":["go"]}: not found

[alecthomas]

I think I saw this before and deleting and recreating the DB container fixed it. See if that works.