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
I'm doing a benchmark of my API ( endpoints are logging traffic and sending them to CloudWatch ) and the amount of memory used increase. After benchmark test finish I see that API still continue working sending logs and seems this task almost never ends.
Seems to me that the Queue used by this module is accumulating logs or objects keeping the node single thread busy trying to purge the queue. I just noticed this module use bottleneck module as a dependency.
I think winston-aws-cloudwatch is not performing well when your API receives tons of requests.
The text was updated successfully, but these errors were encountered:
The behavior should be that it periodically flushes the queue. Of course, if you throw a lot of data at it, it's still going to accumulate that in memory. If it then gets throttled by the CloudWatch API as well, memory is going to get saturated. That said, once items have been flushed from the queue, they should be marked for garbage collection. Thus, as long as flushing is able to keep up with how much data you put in, I see no reason why it would leak.
winston-aws-cloudwatch/lib/queue.js
Line 7 in 7ae96c6
I'm doing a benchmark of my API ( endpoints are logging traffic and sending them to CloudWatch ) and the amount of memory used increase. After benchmark test finish I see that API still continue working sending logs and seems this task almost never ends.
Seems to me that the Queue used by this module is accumulating logs or objects keeping the node single thread busy trying to purge the queue. I just noticed this module use
bottleneck
module as a dependency.I think
winston-aws-cloudwatch
is not performing well when your API receives tons of requests.The text was updated successfully, but these errors were encountered: