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
{{ message }}
This repository has been archived by the owner on Feb 2, 2021. It is now read-only.
I conducted few load test against the plugin and noticed that logs are missing. I set up a job to deploy a container that writes messages to stdout at preconfigured rate, say, 1000 messages per second and the job ran for 10s. At the end of the job, the container will report total message then I compared what we see on blob storage and I noticed there were discrepancy sometimes:
fluentd-load-test-0-td8wm expected: 10001, actual: 10000 x
fluentd-load-test-0-vwtxc expected: 30001, actual: 29998 x
Fluentd container log didn't report any warning and errors during those load tests.
We are using v.11 Fluentd build and I have the exact set up on aws environment but with a different output plugin (s3 output) and I don't see problems there with the same load tests.
Are there parameters I need to tune differently for azure-storage-append-blob plugin to not drop messages?
@dontreboot is it possible to run your load test with 2 outputs? Output logs to some local file and to Append Blob at the same time and compare the results (I believe this allows you to do this: https://docs.fluentd.org/output/copy)
I was running load test using sample source (https://docs.fluentd.org/input/sample) with the rate of 10k messages\second.
It was running for 30 minutes, produced 1.3 GB of logs and had no missing entries (plugin allows you to specify auto_increment_key, so you can see how many messages were posted).
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
I conducted few load test against the plugin and noticed that logs are missing. I set up a job to deploy a container that writes messages to stdout at preconfigured rate, say, 1000 messages per second and the job ran for 10s. At the end of the job, the container will report total message then I compared what we see on blob storage and I noticed there were discrepancy sometimes:
Fluentd container log didn't report any warning and errors during those load tests.
We are using v.11 Fluentd build and I have the exact set up on aws environment but with a different output plugin (s3 output) and I don't see problems there with the same load tests.
Are there parameters I need to tune differently for azure-storage-append-blob plugin to not drop messages?
Here are my configs:
The text was updated successfully, but these errors were encountered: