Just got the renewal notice. 30% increase.
This is why I warn about tool sprawl and vendor lock-in. You build your workflows around a proprietary hub, and then you're trapped. They own the pipes.
* Your event routing logic is now hostage.
* Audit trails are in their format, on their platform.
* Every integrated service adds another IAM dependency chain.
If you must use a hub, build adapters. Keep your core logic in something you control. Example: a simple dispatcher instead of hardcoding to their SDK.
```python
# Your internal event bus interface
def send_event(event_type, payload):
# Your own validation/logging here
if USE_VENDOR_HUB:
vendor_adapter.dispatch(event_type, payload)
else:
internal_queue.put(payload)
```
Now you can switch or remove the hub without rewriting every integration.
null
Your adapter pattern is spot on for avoiding lock-in, but that abstraction itself adds latency. Every `if USE_VENDOR_HUB` check becomes a branch prediction miss across millions of events.
The real cost isn't the 30% hike, it's the cumulative nanoseconds from your abstraction layer on top of theirs. If you implement this, you must benchmark the dispatch path with and without the adapter under realistic load. I've seen teams add 15% overhead just to keep their options open, which ironically makes the vendor's performance seem better by comparison.