Hi,
I'm running Filebeats 5.6.1 and logstash 2.2 on Windows server 2012.
(This was all setup by somebody who has left the company)
We use filebeats to collect the IIS http request log files on a 4 webservers. which go to a single logstash process, then to graphite and some of the metrics go off to elasticsearch.
When we restart the filebeat service on one of the servers, it takes around 4 hours for the metrics to start coming out of the logstash pipeline again. The rest of the metrics carry on coming out of logstash from the other servers as expected.
The logfiles for both Logstash and Filebeat have nothing interesting in them.
could I get some guidence on how to better uncover the delay?
There are two prospectors in the filebeat config, here is the top one:
filebeat:
  prospectors:
    -
      paths:
        - d:\IISLogs\Logfiles\W3SVC1294154473\*.log
      input_type: log
      fields:
        log_source: IIS
        platform: production
        app: sharepoint
        webapp: worksites
      fields_under_root: true
      ignore_older: 2h
      tail_files: true
    -
      paths:
        - d:\IISLogs\Logfiles\W3SVC562970453\*.log
      input_type: log
      fields:
        log_source: IIS
        platform: production
        app: sharepoint
        webapp: quantum
      fields_under_root: true
      ignore_older: 2h
      tail_files: true
  registry_file: "C:/ProgramData/filebeat/registry"
output:
  logstash: 
    hosts: ["xxx.xxx.com:5088"] # hidden
############################# Shipper #########################################
shipper:
############################# Logging #########################################
logging:
  files:
    path: logs
    rotateeverybytes: 10485760 # = 10MB
    keepfiles: 4