Docker Community Forums

Share and learn in the Docker community.

Node.js app in docker container encountered weird dns.lookup latency

(Docker4labs) #1

I was trying to migrate node.js application into docker containers.
When I did some API load test onto our applications which ran in docker containers, I found some API response time were beyond 10 seconds in New Relic.

By looking into the detail, I found some weird things: No matter how many times dns.lookup (external API calls) happened, the dns.lookup took 10,000ms.

I tried to do the same load test to node.js applications running directly on linux servers, I didn’t see any obvious dns.lookup latency.

I didn’t get any clue from Google search results, hope I can get some insight and help here.

Below are 4 New Relic transaction traces:
I’m not allowed to post link. So please copy&paste the string into your browser.

(Iteamnetworkdc) #4

@docker4labs Did you ever get this resolved. Having the same issues with recent Docker Cloud deployments on Node 6.0.0

(Bernaz7) #5

@docker4labs @iteamnetworkdc Has any of you found a resolution for this issue ?
I’m experiencing the exact issue with Node 6.2.0 running on Openshift Enterprise 3.

I think I’ve seen a similar issue regarding Flask as well here :

I can’t have every developer in our environment writing their apps like they did in the above link using Flask (moreover people don’t even think about it because they don’t know it at all)

Is there a good solution for this issue ?

(Iteamnetworkdc) #6

@bernaz7 No. Still seeing intermittent dns.lookup times ranging 200-1000ms. It’s brutal.

(Iteamnetworkdc) #7

Has anyone else seen this issue and might have any ideas. STILL seeing this between micro-service lookups.

(Rizqme) #8

We also have this issue. Is there any solution to this? Please help