Scalable Syslog 12

• Remove Debug Logging

Scalable Syslog 11

Changelog:

• Scheduler consumes adapter via bosh links or a list of static IPs.
• Scheduler no longer attempts to schedule drains on stopped adapters.
• Improve syslog drain retry logic.
• Write syslog errors into applications log stream.
• Write log into application log stream when adapter drops messages.
• Ensure source instance is present in syslog messages.

Scalable Syslog 10

Changelog

• Fix memory leak in adapter when removing a binding that drains to a non existent syslog drain.

Scalable Syslog 9

Changelog:
Fix smoke test

Scalable Syslog 8

Changelog:

• Reduce lock scope in TCP Writer
• Use go-loggregator's pulse emitter to write metrics

Scalable Syslog 7

Changelog:
Increase buffer size in adapter

Scalable Syslog 6

Note: This release fixes a bug with scheduler identified in scalable syslog 5.

Changelog:

• Adapters drain all messages in memory before being bosh stopped. This gives messages an opportunity to reach their final drain destination. The 60s ensures a best effort so adapter will eventually shut down. #142987669
• Added timeouts to ensure that opening streams from adapters to RLP don't hang indefinitely.

Scalable Syslog 5

Known Issue: The above Fix introduced a bug in the adapters that meant streams were not able to be established and drain traffic failed for deployments with large numbers of drains.

Changelog:

• Fixes bug where scheduler didn't clean up extra duplicate drains from adapter. See #145642093

Scalable Syslog 4

Changelog:

• Fix bug that would delay the sending of messages for up to 20 minutes
• Fix slow goroutine leak due to metric reporting
• Fix bug where the scheduler would send a drain to all available adapters. It now assigns a drain to at most two adapters.
• Include instance indexes on logs coming from all components. Example, [CELL/0], [STG/0]...
• Prevent scheduler from deadlocking when unable to communicate with metron or adapters via gRPC

Scalable Syslog 3

Changelog:

• Use gauge metric in scheduler for drain count
• Adapter reconnects to log provider after DNS failure
• Adapter sends filter request for logs only