Big Graphs, No Breakage
tl;dr: Exploring large drip graphs no longer dies with a 400 — traversals now run in batches and just work.
If you've ever tried to explore a really big drip graph and got smacked with a 400 Bad Request, that wasn't you doing anything wrong. Every drip ID in the traversal was being crammed into a single request URL — twice, once for each side of an edge — and past a certain graph size that request simply got too long to send. The bigger your graph, the more likely it was to fail. Which is exactly backwards from how it should work.
Edge lookups now run in batches, with overlapping results deduped on the way back. Traversals return the same edges, in the same order, as before — the only difference is that the large ones actually finish. No new flags, no settings to flip, nothing to migrate. If you hit this error, it's gone.
The practical upshot: graph exploration scales with your work instead of quietly capping out on it. Teams with dense, long-lived project histories — the ones with the most to gain from tracing lineage across hundreds of connected drips — were the ones hitting the ceiling hardest. That ceiling's been raised out of sight.
Same output, none of the errors. Go traverse something enormous.