search

fluent-plugins-nursery / fluent-plugin-cloudwatch-logs

CloudWatch Logs Plugin for Fluentd
MIT License
201 stars 141 forks source link

in_cloudwatch_logs missing logs when start/end specified #234

Closed codertao closed 3 years ago

codertao commented 3 years ago

Problem

Due to other problems, we tried running the cloudwatch_logs input plugin over a specific start/end timeframe to get some missing logs.

When we did that, we only got the most recent bits of logs.

I looked into it, and I think the problem is down to a missing option on the get_log_events API. From the the notes on start-from-head

--start-from-head | --no-start-from-head (boolean)

    If the value is true, the earliest log events are returned first. If the value is false, the latest log events are returned first. The default value is false.

    If you are using nextToken in this operation, you must specify true for startFromHead .

Adding start-from-head=true seemed to fix our issues. And, from the wording, it sounds like it should be used whenever nextToken is present, so might be related to #158 ?

Steps to replicate

 <source>
  @type cloudwatch_logs
  tag a_tag.etc-
  log_group_name /aws/containerinsights/removed
  log_stream_name removed-
  use_log_stream_name_prefix true
  region us-east-1
  start_time "2020-05-01 00:00:00"
  end_time "2020-05-13 23:59:59"
  use_aws_timestamp true
  <storage>
    @type local
    path /data/state.json
  </storage>
</source>
<match a_tag.**>
   @type file
   path /data/logs
   <buffer>
     @type memory
     timekey 3600
     chunk_limit_size 1M
     flush_at_shutdown true
   </buffer>
 </match>

On running, we get a few thousand log lines from the end of the time range... and also logs from today, for some reason, and none of the older logs in the time range.

Expected Behavior or What you need to ask

Should get all the logs from the time range

Using Fluentd and CloudWatchLogs plugin versions

Nix OS, running in Docker via fluent/fluentd:v1.7-1 image.

/ $ fluentd --version fluentd 1.12.3

/ $ fluent-gem list

LOCAL GEMS

async (1.23.0) async-http (0.46.3) async-io (1.27.0) aws-eventstream (1.1.1) aws-partitions (1.452.0) aws-sdk-cloudwatchlogs (1.40.0) aws-sdk-core (3.114.0) aws-sigv4 (1.2.3) bigdecimal (1.3.5) bundler (2.2.17) cmath (default: 1.0.0) concurrent-ruby (1.1.5) console (1.6.0) cool.io (1.5.4) csv (default: 1.0.0) date (default: 1.0.0) dig_rb (1.0.1) elasticsearch (7.12.0) elasticsearch-api (7.12.0) elasticsearch-transport (7.12.0) etc (default: 1.0.0) excon (0.81.0) faraday (1.4.1) faraday-excon (1.1.0) faraday-net_http (1.0.1) faraday-net_http_persistent (1.1.0) fcntl (default: 1.0.0) fileutils (default: 1.0.2) fluent-config-regexp-type (1.0.0) fluent-plugin-cloudwatch-logs (0.13.4) fluent-plugin-concat (2.4.0) fluent-plugin-elasticsearch (5.0.3) fluent-plugin-filter_typecast (0.0.3) fluent-plugin-grok-parser (2.6.2) fluent-plugin-prometheus (2.0.1) fluent-plugin-rewrite-tag-filter (2.4.0) fluent-plugin-route (1.0.0) fluentd (1.12.3, 1.7.4) http_parser.rb (0.6.0) ipaddr (default: 1.2.0) jmespath (1.4.0) json (2.2.0) msgpack (1.3.1) multi_json (1.15.0) multipart-post (2.1.1) nio4r (2.5.2) oj (3.3.10) openssl (default: 2.1.2) prometheus-client (2.1.0) protocol-hpack (1.4.1) protocol-http (0.8.1) protocol-http1 (0.8.3) protocol-http2 (0.9.7) psych (default: 3.0.2) ruby2_keywords (0.0.4) scanf (default: 1.0.0) serverengine (2.2.3, 2.2.0) sigdump (0.2.4) stringio (default: 0.0.1) strptime (0.2.3) strscan (default: 1.0.0) timers (4.3.0) tzinfo (2.0.0) tzinfo-data (1.2019.3) webrick (default: 1.4.2) yajl-ruby (1.4.1) zlib (default: 1.0.0)

cosmo0920 commented 3 years ago

This should be done by https://github.com/fluent-plugins-nursery/fluent-plugin-cloudwatch-logs/pull/235. Closing.