Conversation
|
Finally got around to fixing this up, ready for review! |
|
One thing I'd like to know first is why this PR seems to revert commit 55587d1 and implement the Generation IDs suddenly? This does not seem necessary at all. |
We're using the Generation ID's pretty heavily as it allows for us to track propagation throughout the fleet of servers. It allows us to keep track of how long it takes for the snapshots to propagate. |
|
I've discussed this internally and I'm not comfortable pulling this in its current state. The (re)introduction of GenerationID with a very different purpose than what it was originally used for, plus the numerous metrics changes are thing I'd rather not review for the purpose of including Azure support in Lightning Stream. I'll have to account for the interactions with our enterprise version as well, so a more limited PR with just the storage parts is more likely to get merged more quickly. I'm happy to discuss any issues tracking snapshot propagation, or metrics you'd like to see added. You're more than welcome to open a few Discussions. |
|
Closing in favor of #121 |
This pull request adds support for Azure Blob Storage as a backend for storing snapshots in Lightning Stream, alongside existing S3 and filesystem options. It includes documentation updates, configuration examples, and local testing support via Azurite in Docker Compose. The changes also update environment variable formatting for Docker services and clarify sweeper configuration.
cc @Luit @ahouene