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
Hi,
it seems logspout is not able to capture all logs produced by docker. From time to time some log lines are missing in LogDNA.
If I choose INACTIVITY_TIMEOUT=5s all logs make it through, but every ~5s the CPU usage spikes at 100% due to docker parsing all logfiles (see gliderlabs/logspout#246).
Setting TAIL=0 is resulting in lower CPU usage, but some log lines are missing.
All containers are running on a Docker Swarm cluster.
Docker: 18.09.1
Docker log-driver: json-file
logspout: v3.2.6-custom (https://hub.docker.com/r/logdna/logspout)
Do you hav any idea how to properly send all log lines without risking high CPU usage? Or is there another way to get Docker Swarm container logs to LogDNA?
Kind regards,
Maxi
The text was updated successfully, but these errors were encountered:
Hi,
it seems logspout is not able to capture all logs produced by docker. From time to time some log lines are missing in LogDNA.
If I choose INACTIVITY_TIMEOUT=5s all logs make it through, but every ~5s the CPU usage spikes at 100% due to docker parsing all logfiles (see gliderlabs/logspout#246).
Setting TAIL=0 is resulting in lower CPU usage, but some log lines are missing.
All containers are running on a Docker Swarm cluster.
Docker: 18.09.1
Docker log-driver: json-file
logspout: v3.2.6-custom (https://hub.docker.com/r/logdna/logspout)
Do you hav any idea how to properly send all log lines without risking high CPU usage? Or is there another way to get Docker Swarm container logs to LogDNA?
Kind regards,
Maxi
The text was updated successfully, but these errors were encountered: