Request timed out. Retries due to rate limiting: True.

dmitrii sablin 0

I use azure cosmos db for mongo db (4.2 version) with 10.000 Ru/s
a have a collection with 2.8m of records. the field "type" are indexed via a Single field index.
when i am trying to perform the query:

db.Actions.aggregate([ {"$group" : {_id:"$type"}} ])

i receive the message:

{
        "ok" : 0,
        "errmsg" : "Request timed out. Retries due to rate limiting: True.",
        "code" : 50,
        "codeName" : "ExceededTimeLimit"
} : aggregate failed

dmitrii sablin 0 Reputation points

2023-05-14T12:07:13.21+00:00

i also can not understand why, whem im using mongoDB Driver, db.collection.Find(<any_filter>).Count() works realy faster than any db.collection.CountDocuments(<any_filter>) ??_
dmitrii sablin 0 Reputation points

2023-05-14T13:15:59.5066667+00:00

@Konstantinos Passadis SSR already enabled.

I just can not understand, why 10.000 Ru/s is not enough just for the grouping result?
dmitrii sablin 0 Reputation points

2023-05-14T13:18:32.99+00:00

@Konstantinos Passadis SSR already enabled. But i use 10.000 Ru/s. Is that really not enought just for the groupping the results?
Also please see my second comment. I still can not understand why find.count works faster than countDocumentsAsync
Konstantinos Passadis 17,276 Reputation points

2023-05-14T14:28:27.7233333+00:00

Hello @dmitrii sablin !

I can understand your questions , and the whys are also dependable

For example a db.Actions.aggregate([ {"$group" : {_id:"$type"}} ])

The $group stage separates documents into groups according to a "group key". The output is one document for each unique group key. This can be resource-intensive as it goes through all documents thus getting more RUs

Now regarding find.count and countDocuments , the first one is returning the total number of results that match a query , but the second one returns the number of documents in the collection that match the query. it is logical the first one is faster !

There is a chance that you may need to redesign the Project , consider alternatives or use combinations like Synapse Analytics , or Data Lake Storage

I hope this helps!

Kindly mark the answer as Accepted and Upvote in case it helped!

Regards
dmitrii sablin 0 Reputation points

2023-05-14T15:58:26.9966667+00:00

Sorry, but what is the difference between “total number of results that match a query” and

“number of documents in the collection that match the query”

sounds the same 🤷🏻
Konstantinos Passadis 17,276 Reputation points

2023-05-14T16:15:54.1133333+00:00

Hello @dmitrii sablin !

I think it is more of how each one works

Reading this

https://stackoverflow.com/questions/62998919/mongodb-countdocuments-vs-exists-vs-find-vs-findone-which-is-the-fastest-way-to

And being through the Documentantion

https://www.mongodb.com/docs/manual/reference/method/db.collection.find/

My understanding is that db.collection.count()

Returns the count of documents that would match a find() query for the collection or view. The db.collection.count() method does not perform the find() operation but instead counts and returns the number of results that match a query.

I hope this helps!

Kindly mark the answer as Accepted and Upvote in case it helped!

Regards
ShaktiSingh-MSFT 13,271 Reputation points Microsoft Employee

2023-05-16T12:02:46.0466667+00:00

Hi dmitrii sablin •,

Welcome to Microsoft Q&A forum.

For this issue of Request timed out. Retries due to rate limiting: True, if you are still stuck and could not enable SSR, I recommend you to open a support ticket to get server-side retries turned on.

Please let me know if further queries, thanks

1 answer

Konstantinos Passadis 17,276 Reputation points

2023-05-14T12:13:17.81+00:00

Hello @dmitrii sablin !

I understand that your CosmosDB with MongoAPI is getting rate limited .

Azure Cosmos DB for MongoDB operations may fail with rate-limiting (16500/429) errors if they exceed a collection's throughput limit (RUs).

You can enable the Server Side Retry (SSR) feature and let the server retry these operations automatically. The requests are retried after a short delay for all collections in your account. This feature is a convenient alternative to handling rate-limiting errors in the client application.

Have a good look here :

https://learn.microsoft.com/en-us/azure/cosmos-db/mongodb/prevent-rate-limiting-errors

You could also utilize Change Feed , and Increase Provisioned Throughput | Request Units Change

I hope this helps!

Kindly mark the answer as Accepted and Upvote in case it helped!

Regards
Please sign in to rate this answer.

0 comments No comments
Sign in to comment