Score:2

Are the units incorrect for Storage Read / Write graphs in AWS ECS Container Insights?

in flag

In AWS Console > CloudWatch > Container Insights > Performance monitoring > ECS Tasks, the graphs for both Storage Read and Storage Write have units shown as Bytes/Second like this:

Storage Read

Bytes/Second

However in both cases the graphs appear to be always monotonically increasing, except for vertical drops on container restart, and have long horizontal sections when I would not expect to be seeing sustained constant IO.

Are these graphs actually showing cumulative Bytes rather than the stated Bytes/Second?

Looking in the table on AWS docs for ECS Container Insights metrics I see StorageReadBytes and StorageWriteBytes with both say Unit: Bytes so, assuming this is referring to the same metrics displayed in the console UI graphs, perhaps this confirms my suspicion that this is a bug in the AWS console UI?

I have attempted to report this using the 'Feedback' button in AWS but posting here in case anyone can confirm or if it helps others who may be worried by a seemingly high rate of sustained I/O.

Nicholas Smith avatar
us flag
I just ran into the same and found your question here. The CloudWatch metric shows bytes/sec but clearly my application is not writing 20GB/sec (for example).. it seems to be a sum.
sa flag
This still seems to be an an issue. According to Storage Write on container insight one of our tasks are now writing at >100 GB/sec - and increasing!
Score:0
cn flag

Yes, it seems the units are incorrect. I also ran into same observation while debugging an application restart. It took a while. For me the value was 215 mb/s for StorageReadBytes - when the application was idle. I think this needs to be raised to AWS team.

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.