Long Cold-Start Times on US East 2
I'm seeing some very long cold start times on my free-tier instance (project id: weathered-voice-16568128). Performance is great after the first read, but that read can take 13s (as measured by the Neon Console operations time).
I'm using the free-tier now leading up to the launch of a product I'm working on. Before the launch, I plan to upgrade to the Launch tier. As described here, I don't see any mention of scaling control or cold-start priority.
I wanted to write this in case there's something worth looking into. I also have two questions:
1. Should I expect to see these cold start times even after upgrading to the Launch tier?
2. Is there a way to pipe database metrics (query time, cold start times) somewhere? I'm on AWS, so I'd likily fire off a CloudWatch alert.
For (2), it seems like cursor-based polling is the only option today. Is that correct? If so, it would be great to have a webhook for important events.
I'm using the free-tier now leading up to the launch of a product I'm working on. Before the launch, I plan to upgrade to the Launch tier. As described here, I don't see any mention of scaling control or cold-start priority.
I wanted to write this in case there's something worth looking into. I also have two questions:
1. Should I expect to see these cold start times even after upgrading to the Launch tier?
2. Is there a way to pipe database metrics (query time, cold start times) somewhere? I'm on AWS, so I'd likily fire off a CloudWatch alert.
For (2), it seems like cursor-based polling is the only option today. Is that correct? If so, it would be great to have a webhook for important events.
Neon
Retrieves a list of operations for the specified Neon project.You can obtain a project_id by listing the projects for your Neon account.The number of operations returned can be large.To paginate the response, issue an initial request with a limit value.Then, add the cursor value that was returned in...
