[COMPASS-4504] MongoDB Compass cause Atlas Disk I/O % above 90% Created: 18/Nov/20  Updated: 08/Aug/22

Status: Open
Project: Compass
Component/s: Atlas, Performance
Affects Version/s: 1.23.0
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Yang Liu Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

macOS 10.15.7, M40 with 3 shards


Attachments: PNG File image-2021-02-02-00-15-11-284.png    
Story Points: 5

 Description   

Problem Description

Whenever I have MongoDB Compass connect to our live database, after couple minutes I'll receive Atlas alert "Disk I/O % utilization on Data Partition has gone above 90". And the alert keeps triggering while I keep Compass open. 

Steps to Reproduce

(My best guess, not minimal reproducible.)

  1. Create a MongoDB Atlas cluster, M40 with 3 shards.
  2. Fill some data to a collection and make it around 25 million documents and 50GB. 
  3. Have MongoDB Compass connect to the cluster. Select that collection and show preview data. 

Expected Results

There should not be alert triggered. 

Actual Results

Atlas alert "Disk I/O % utilization on Data Partition has gone above 90"

Additional Notes



 Comments   
Comment by Yang Liu [ 02/Feb/21 ]

Hi, thanks for looking into this and I'm really sorry for the late reply. 

I run each command for about 5 minutes, and didn't get alerts or seeing any I/O peak in the Atlas "Realtime" monitor.

When I use Compass to show the same collection, I does show the alerts. (But I cannot catch a obvious peak in the "Realtime" monitor either)

The collection size being tested is as below

 

 

Comment by Massimiliano Marcon [ 18/Nov/20 ]

yang@thatgamecompany.com can you connect with the shell and run:

db.collection.countDocuments()


and

db.collection.aggregate( [
 { $count: "myCount" }
])

(replacing collection with the name of your sharded collection.

and check if in both cases you see the same spike in I/O you saw with Compass?

Comment by Yang Liu [ 18/Nov/20 ]

As a note, we had a consultant session with Karan from MongoDB who witnessed the issue we had. 

Generated at Wed Feb 07 22:36:35 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.