You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using syslog-ng in a hub and spoke pattern it is desirable to scale up and own the number of instances based on throughput needs. When using memory queues and excluding file queues and when no other stateful needs such as python check points syslog-ng should be able to handle a stop signal to shut down in a graceful way with no data loss caused.
Proposed solution
The term signal is received
Source models as closed/terminated, it is important to note how this is done is on a case by case basis and may require source side enhancement. For the initial release syslog,network,file,named-pipe,program, journald,otel,http and files would cover most instances. A configurable shut down time for this phase is desirable
After all sources have closed monitor the queue depth of destinations until each reaches zero or a timer expires. The total shutdown timer should begin at the time the signal is received
Alternatives
External proxy, connection management to drain considered to difficult.
Additional context
The text was updated successfully, but these errors were encountered:
Description of the problem
When using syslog-ng in a hub and spoke pattern it is desirable to scale up and own the number of instances based on throughput needs. When using memory queues and excluding file queues and when no other stateful needs such as python check points syslog-ng should be able to handle a stop signal to shut down in a graceful way with no data loss caused.
Proposed solution
Alternatives
External proxy, connection management to drain considered to difficult.
Additional context
The text was updated successfully, but these errors were encountered: