Database
301 results found
-
Allow to decrease time series granularity and custom bucketing values
In our IoT use case, we are leveraging MongoDB’s time series functionality. Due to high write volume, we need to adjust the timeseries.granularity and bucketMaxSpanSeconds parameters to manage the write load. However, after increasing the bucketMaxSpanSeconds, we need to run the system for several days to observe stability. If the value is set too high, MongoDB does not support decreasing the bucketing value, and we are forced to create a new collection instead.
It would greatly simplify our testing process and increase flexibility in adapting to business changes if MongoDB allowed the decrease of bucketMaxSpanSeconds after it has been increased.
2 votes -
Metadata (Created Date) for database accounts
We frequently have audit questions about when a given user was created in our Mongo databases. Other DBMSs (Oracle, for instance) have a 'CREATED' field in a metadata table (i.e. DBA_USERS) that shows when a user was created, and this would be very helpful when responding to these audits.
2 votes -
Making operations that hang indefinitely more visible
Hi Team,
It would be useful for some kind of timeout message to be implemented if an operation is waiting on commitQuorum - say wait an hour and then return an error message with "Can't perform operation, required number of commitQuorum members not available to process this operation" or something similar. I don't think a command should hang indefinitely, and it can also help both customers and support for the analysis of this kind of issues.
Thanks in advance
2 votes -
log redaction customization
If we have customizations on the log redaction feature then we can add our preference where we have to apply log redaction. As of now according to my opinion it is getting applied on the all log fields which may create a issue while troubleshooting the any issue.
2 votes -
Named MongoDB Connections
When a service is acting erroneously and generating hundreds or thousands of connections, it's currently difficult to determine which service is doing so when you have 30+ services connecting to MongoDB.
My proposal is that we should be optionally able to specify a non-unique name for the connection in the MongoDB URL (possibly after a #), which would allow DB administrators to see how many of each named connection was connected at any given time, and also other metrics (operations/s per name, average/max query execution time per name, etc.).
Example URL:
mongodb://user:****@clustername.abcd.mongodb.net:27017/dbName?authSource=admin#myServiceName2 votes -
Providing connection details with username for real time connection monitoring.
Currently, mongd.log provides connection details like remote IP, source port with authentication information but doesn't provide the connection is in active state or not.
serverStatus() only provides number of current and active connections.Example:
10.0.0.100:12345 - username active
10.0.0.101:12346 - username idle2 votes -
Export Backup Snapshots to GCP Bucket
As Atlas user, i want to Export Backup Snapshots to a GCP bucket because i don't have an AWS subscription.
2 votes -
Make $merge support DELETE operation
Currently, the
$mergeonly supports insert/update/upsert/merge behaviour. It would be great to support delete behaviour. A common use case would be in-place document deduplication/clean-up in a collection.2 votes -
Change Timeseries Bucket Memory Limit
Currently, there's no way to set a memory threshold for bucket allocation in MongoDB's time series collections. As the bucket size increases and more collections are opened day after day, a limiting mechanism triggers for open buckets, leading to cache pressure and the premature closing of buckets under high load. It would be beneficial for users to have the ability to set a memory threshold for the timeseries bucket memory limit ( I guess limit is around 3GB). It enabling us to prevent early bucket closure in production environments. Alternatively, providing the option to manually close buckets could help manage…
2 votes -
Include or flag nodes seen more than once / duplicates when traversing with $graphLookup
When traversing a graph with $graphLookup, we suppress duplicates (in cases of cycles). It would be useful to have an option to include a flag indicating a node that caused the traversal to stop due to duplication.
2 votes -
TTL index throttling for deletion
To avoid deletion impact, we need to control the quantity and speed of document generations(insert).
Need a way to throttle down the deletion quantity per a delete operation.i.e. # of documents per a ttl delete, sleep time between deletions(default 60s)
2 votes -
Improve Changestream metadata by including userId(who), action/intent (event name).
This can be done by including meta data in the options for C(R)UD operations!
Something like:User.updateOne({ _id }, { $set: { name: "newName" } }, { $meta: { userId: _id, action: "nameUpdated" })This is super useful because now changestream can be used to create event store or audit logs, out of the box. I just have to store all the changestream events from the relevant collections. This also makes event driven architecture in micro services super useful. Developers can use this to publish events directly to a message broker such as kafka, where action would be equivalent to…
2 votes -
I just wanted to say Good job!!!
Appreciation
2 votes -
define the random seed manually, for $rand and $sample
It will be great if an additional paramater to define the seed for $rand and $sample could be use.
2 votes -
Tool for score data model
It would be great to have a tool which can scores a data model from a specific database. I mean, I could allows this tool to scan a database model and score this data model based on best practices, patterns and anti patterns. It could also generates a list of problems and suggestions for improvement.
This tool could be used on CI pipelines near to the unit tests and the main goal is to avoid release new features using a bad data model.
2 votes -
Flatten arrays in group stage
Have group operators to flatten document arrays into a single one with or without repeated elements.
So ->
doc1 = {arr: [1,2,3,4], gr: "group"}, doc2 = {arr: [5, 6, 7, 8], gr: "group"}
{$group: {id: "$gr", arrays: {$***: "$arr"} } }
=>
{id: "gr", arrays: [1, 2, 3, 4, 5, 6, 7, 8]}2 votes -
Cascading delete for DBRefs
Since transactions have been added in 2018, which work across collections (https://www.mongodb.com/docs/manual/core/transactions/) and across shards (https://www.mongodb.com/docs/manual/core/transactions-sharded-clusters/), shouldn't cascading deletes be possible now? I only worked with sql transactions in the past, but my intuition would be that it should be fairly easy to do this in a client:
- start a transaction
- fetch document
- look for dbref fields
- fetch those docs
- continue at 2 until all docs have been found, stopping at branches when a doc has already been fetched
- go back in reverse and delete all of them
- commit transaction
If this is possible to do…
2 votes -
geo
It would be nice to get the length of an LineString of a geo-json object or the possibility to write an aggregation to calculate it.
2 votes -
Support Two Array Fields in Compound Indexes
Hi,
I've come across a lot of use cases where the business logic has demanded unique constraints on 1-2 fields that have been modeled as arrays on documents. The cases that only contain a single array field is already taken care of using unique indexes in MongoDB, however the cases with 2 array fields have required application level constraints since the database only supports a single array field in compound indexes. While multiple arrays in a single index would massively increase the size of the indexes, it would be very helpful if it would still be possible. If there are…
2 votes -
Query Planner needs a timeOut set as a Database parameter.
Query Planner needs a timeOut set as a Database parameter.
We see App queries timing out on Query planner taking >1 sec. Whilst this can be avoided by setting maxTimeMS at client end - this is more of a setting for the overall query and not just the query planner - Also, this comes at the risk of closing/timing out the actual query (cursor) which is not our need.
We only want the query planner itself to have a specific - set/customisable timeout and the query to keep running selecting one/any if the plans run thus far without timing out…
2 votes
- Don't see your idea?