I was hoping this was fixed in https://github.com/logstash-plugins/logstash-filter-memcached/pull/25 with the plugin release v1.1.0
Related issue was already raised in Catch Memcached Filter exceptions but this thread is closed so I am crating a new one.
Doing a simple test starting LS with memcached host unavailable. Getting message below:
[2020-11-03T03:04:23,322][ERROR][logstash.filters.memcached][sysmon_servers] failed to connect to memcached {:hosts=>["memcached:11211"], :options=>{:expires_in=>0}, :message=>"No server available"}
[2020-11-03T03:04:23,321][ERROR][logstash.filters.memcached][security_servers] failed to connect to memcached {:hosts=>["memcached:11211"], :options=>{:expires_in=>1800}, :message=>"No server available"}
[2020-11-03T03:04:23,346][ERROR][logstash.agent ] Failed to execute action {:id=>:sysmon_servers, :action_type=>LogStash::ConvergeResult::FailedAction, :message=>"Could not execute action: PipelineAction::Create<sysmon_servers>, action_result: false", :backtrace=>nil}
[2020-11-03T03:04:23,348][ERROR][logstash.agent ] Failed to execute action {:id=>:security_servers, :action_type=>LogStash::ConvergeResult::FailedAction, :message=>"Could not execute action: PipelineAction::Create<security_servers>, action_result: false", :backtrace=>nil}
[2020-11-03T03:28:06,823][WARN ][org.logstash.plugins.pipeline.PipelineBus][distributor][6360b3d362cbbe9487b6ad8170404b906cf2f481c1aab51aa68c344b8b1ed2aa] Attempted to send event to 'security_servers' but that address was unavailable. Maybe the destination pipeline is down or stopping? Will Retry.
[2020-11-03T03:28:07,824][WARN ][org.logstash.plugins.pipeline.PipelineBus][distributor][6360b3d362cbbe9487b6ad8170404b906cf2f481c1aab51aa68c344b8b1ed2aa] Attempted to send event to 'security_servers' but that address was unavailable. Maybe the destination pipeline is down or stopping? Will Retry.
Manually starting memcached service. Despite memcached is now available pipelines never recover until I manually restart Logstash, which is not ideal. Did PR mean to fix described behaviour? Is there possibly a way to handle it better?
TA
A