https://community.splunk.com/t5/Getting-Data-In/TCP-Buffer-Options/m-p/214707#M42253 <P>We are using a Cloud Foundry for an Internal Cloud Implementation. We are migrating applications and hence using TCP streaming versus the Splunk Forwarding Agent to stream logs. The Cloud Support Team is observing errors logged because the TCP stream is backing up. I am using a vanilla TCP configuration. Are there extra configurations I can add to the stanza to increase the buffer to prevent messages from dropping?</P> <P>Sample Vanilla TCP configuration:<BR /> [tcp://3301]<BR /> connection_host = dns<BR /> index = index_name<BR /> sourcetype = rfc5424_syslog</P> <P>Logs from the Cloud Foundry Doppler Service:<BR /> {"timestamp":1474911116.219854593,"process_id":17528,"source":"doppler","log_level":"warn","message":"TB: Output channel too full","data":{"appId":"d41c5c78-955a-4148-b575-cf868dc0b6fe","destination":"syslog://tlaloga1.dev.prod.travp.net:3303","dropped":99,"total_dropped":496},"file":"/var/vcap/data/compile/doppler/loggregator/src/truncatingbuffer/truncating_buffer.go","line":112,"method":"truncatingbuffer.(*TruncatingBuffer).forwardMessage"}<BR /> {"timestamp":1474911142.510904074,"process_id":17528,"source":"doppler","log_level":"warn","message":"TB: Output channel too full","data":{"appId":"d41c5c78-955a-4148-b575-cf868dc0b6fe","destination":"syslog://tlaloga1.dev.prod.travp.net:3303","dropped":99,"total_dropped":595},"file":"/var/vcap/data/compile/doppler/loggregator/src/truncatingbuffer/truncating_buffer.go","line":112,"method":"truncatingbuffer.(*TruncatingBuffer).forwardMessage"}<BR /> {"timestamp":1474911161.138316393,"process_id":17528,"source":"doppler","log_level":"warn","message":"TB: Output channel too </P> Tue, 29 Sep 2020 11:10:15 GMT ezajac 2020-09-29T11:10:15Z https://community.splunk.com/t5/Getting-Data-In/TCP-Buffer-Options/m-p/214707#M42253 <P>We are using a Cloud Foundry for an Internal Cloud Implementation. We are migrating applications and hence using TCP streaming versus the Splunk Forwarding Agent to stream logs. The Cloud Support Team is observing errors logged because the TCP stream is backing up. I am using a vanilla TCP configuration. Are there extra configurations I can add to the stanza to increase the buffer to prevent messages from dropping?</P> <P>Sample Vanilla TCP configuration:<BR /> [tcp://3301]<BR /> connection_host = dns<BR /> index = index_name<BR /> sourcetype = rfc5424_syslog</P> <P>Logs from the Cloud Foundry Doppler Service:<BR /> {"timestamp":1474911116.219854593,"process_id":17528,"source":"doppler","log_level":"warn","message":"TB: Output channel too full","data":{"appId":"d41c5c78-955a-4148-b575-cf868dc0b6fe","destination":"syslog://tlaloga1.dev.prod.travp.net:3303","dropped":99,"total_dropped":496},"file":"/var/vcap/data/compile/doppler/loggregator/src/truncatingbuffer/truncating_buffer.go","line":112,"method":"truncatingbuffer.(*TruncatingBuffer).forwardMessage"}<BR /> {"timestamp":1474911142.510904074,"process_id":17528,"source":"doppler","log_level":"warn","message":"TB: Output channel too full","data":{"appId":"d41c5c78-955a-4148-b575-cf868dc0b6fe","destination":"syslog://tlaloga1.dev.prod.travp.net:3303","dropped":99,"total_dropped":595},"file":"/var/vcap/data/compile/doppler/loggregator/src/truncatingbuffer/truncating_buffer.go","line":112,"method":"truncatingbuffer.(*TruncatingBuffer).forwardMessage"}<BR /> {"timestamp":1474911161.138316393,"process_id":17528,"source":"doppler","log_level":"warn","message":"TB: Output channel too </P> Tue, 29 Sep 2020 11:10:15 GMT https://community.splunk.com/t5/Getting-Data-In/TCP-Buffer-Options/m-p/214707#M42253 ezajac 2020-09-29T11:10:15Z https://community.splunk.com/t5/Getting-Data-In/TCP-Buffer-Options/m-p/214708#M42254 <P>Rather than increase the buffer size you can scale the number of Dopplers available. Below is a link to a guide to how many Dopplers and Traffic Controllers to plan for. </P> <P><A href="https://discuss.pivotal.io/hc/en-us/articles/225564028-How-to-Calculate-the-Loggregators-Message-Throughput,Rather">https://discuss.pivotal.io/hc/en-us/articles/225564028-How-to-Calculate-the-Loggregators-Message-Throughput,Rather</A> than scaling the buffer size the approach we recommend is scaling the number of dopplers. See <A href="https://discuss.pivotal.io/hc/en-us/articles/225564028-How-to-Calculate-the-Loggregators-Message-Throughput">https://discuss.pivotal.io/hc/en-us/articles/225564028-How-to-Calculate-the-Loggregators-Message-Throughput</A> for a guide to scaling more dopplers (and possibly Traffic Controllers) as well. </P> Wed, 30 Nov 2016 18:10:16 GMT https://community.splunk.com/t5/Getting-Data-In/TCP-Buffer-Options/m-p/214708#M42254 ahev 2016-11-30T18:10:16Z