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
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 ?
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.
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
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
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)
The text was updated successfully, but these errors were encountered: