Suggest features, upvote favorites, and use the user story format: > As a [role], I want to [action], so that [benefit]. Help us build what you need - post your idea now!
What you should know submitting product ideas to our product team
Share your product ideas with the community and our product team
Cognite Charts Team, I would like Cognite Charts and CDF help User From Celanese:Here’s my scenario – I’m going to be as vague as an actual problem statement that we would get – At the base of T109 are a set of pumps. I believe they basically pump on-spec AA to the tankfarm. They have been debottlenecked so much over the years that they have essentially run out of NPSH. The scenario is to find the NPSHa of these pumps and figure out how often they cavitate and cause upsets (I’m honestly not sure if they do or not). To do a full analysis on that we’re going to need: The NPSHr of the pumps (pump datasheets aren’t in Bentley, so that’s a problem but I can get it) The historical level in T109 and the ability to translate that level in actual ft of elevation The elevation of T109 relative to the pumps (ie, the isometric on the pumps’ suction or 3D) The temp in T109 The flow from the pumps to the tank farm The ability to correlate drops in flow to the T109 level and temperature Any other time series around this system that may be helpful in the analysis We can then build a calculation showing the NPSHa of those pumps and track it in real time – ideally we would also set an alert when it started to get marginalHow would i get to the last point without speniding hours searching in CDF and then flipping to Charts? Any ideas how i can have all that to build the caclulation in Charts. Access to ISOMETRICS, 3D, Time series and P&ID
From the meeting on 10/22/24.Customers would find Operations in order more useful in Search UX.Can a sort field be added under Operations either by name or by sequence?More than 25 entries will be needed by the customer
From the meeting on 10/22/24Customers need a way to sort in the other areas like equipment. Currently only in assets for sort
Repository:https://github.com/cognitedata/using-cognite-python-sdkDescription:As a developer working with the Cognite Python SDK, I would find it immensely helpful if the repository provided a Dockerfile to set up a virtual environment. This would allow for an environment to be set up consistently and effortlessly with just a few commands ( like "docker build" and "docker run" ) , saving developers from potential dependency issues or misconfigurations.Benefits:Consistent Setup: Eliminates environment discrepancies across different systems, ensuring consistent setup for all users.Time-Saving: Significantly reduces the time developers spend resolving dependency issues.Scalability: Useful for CI/CD pipeline integration in case needed in future.Conclusion:Adding a Dockerfile would make the onboarding process smoother for new users and improve the overall development experience. I'd appreciate if you could consider this feature request!Thank you :)
It is very difficult to search files in Cognite if you dont know the file name. For example, i wanted to search electrical classification drawing for a methanol unit in CDF. it was practically impossible to find it in CDF. Instead, I had to go to the source system → to find the unit methanol and then find the electical classification drawing under electrical drawing folder. There is no way in CDF to search by document classes and/or filter by unit to narrow the search in a few clicks. Also there is no way to isolate fuzzy searches so that i dont get random stuff.
We would like to see the ability to rollup documents by type, using metadata, within the new search UI and Industrial Canvas asset browser, as well as other places where users are looking for documents associated with an asset (similar request to this topic Cognite Hub). We have seen this in the legacy data explorer view and it was well received by our users and our document teams have started to shift their business process to tag documents with the appropriate metadata to enable this functionality.Our assets can have hundreds of different documents associated with them, so an ungrouped list will make it hard for users to quickly find what they are looking for. In addition, we have some use cases where documents are snap-shot and stored as a part of a project file (e.g. P&IDs). Separating these duplicates into the correct rollup buckets will allow us to eliminate confusion on which files are most recent to use. Example for old search:
This is feedback from a unit engineer:When I use the new search experience on Cognite, I noticed the files for a piece of equipment aren’t sorted in to categories unlike the classic data exploration. This is a functionality I really liked in the classic data exploration as it saved time digging through files. Is this a functionality that could be applied to the new search experience?
Filling out on behalf of an engineer super user: When looking through the files of an asset, I only see the P&ID number in the “Tree” view. There should be more information like P&ID title or other details while still in the Tree view so it’s easier to understand what the P&ID contains without having to enter each P&ID/equipment drawing to find out.Scratched the names out in the image below, but say you have a P&ID named “12AB0123” - this doesn’t tell the user very much information.
Add support for sequencing-specific extractors and transformations and downstream objects. This would enable users to easily use tools for common sequencing data processing tasks.Key features:support to trigger data workflows post competition of extraction process.currently it is happening in silos and the transformation assumes that the extraction process is complete and runs on a predefined frequency. we would like to have the transformation triggered post completion of the extraction process Automated Workflows in relativity enables users to streamline and automate data processing tasks, particularly for sequencing-specific operations. Key features include:Sequencing-Specific Extractors and TransformationsThe system should support specialized extractors and transformations tailored for sequencing data. This allows users to easily process common sequencing data types without manual intervention.Downstream Object SupportAutomated Workflows can now handle downstream objects, enabling more complex data processing pipelines that depend on earlier extraction and transformation steps.Triggered TransformationsA crucial improvement is the ability to trigger transformations immediately upon completion of the extraction process. This replaces the previous model where transformations ran on predefined schedules, regardless of extraction status.Workflow ChainingUsers should be able to chain multiple workflows together, with each step triggering the next upon completion. This creates a seamless, end-to-end automated process for data handling.Customizable TriggersThe system offers flexible trigger options, allowing workflows to start based on various conditions such as completion of other workflows.
FDM documentation should have more examples.
Currently, aggregate query options for timestamp and date typed properties are not available in both GraphQL and SDK. Additionally, there is no method to aggregate by a time range, such as monthly or weekly, similar to here available for timeseries.Using groupBy notificationDate (timestamp)Using 'groupBy' with a timestamp/date field, it would be feasible to aggregate data from a data model view by timestamp/date fields, group them by daily granularity, and display the count of a record for each day.Related to this topic.
The tags for Yggdrasil (and other assets) are represented in CDF. There is a need for making relations between tags to build more knowledge into the model.I am working within the maintenance domain, and the simplest use case where such a capability would be needed, is when the CBM engineer want to express that two (or three) sensors are comparable in terms of measurement value. The concept is called transmitter comparison, and it ties into the fact that if two or more transmitters are to output the same value (pressure, temperature, level etc.), you can monitor the outputted values, and perform maintenance if one of the transmitters starts drifting/freezes etc.Currently these kind of relationships are kept in the condition monitoring system, it is difficult to integrate changes to the plant (modifications) that might affect such models. Keeping these relation close to the LCI-data in CDF, would enable governance on this kind of data. CBM engineers currently reads P&IDs to identify relevant transmitter pairs (or trios). When such a relationship is identified, this needs to be expressed in a model (13PT1012 isComparable to 13PST1202). One possible solution would be to add this capability to the Graph with available in DM. There you can filter out relevant tags, and one could imagine being able to “draw lines between tags” to represent the relationship. In the background a transformation could be run on demand, establishing the relationship. As mentioned, this is the simplest use case. If you want to monitor a centrifugal pump, you would like to use approx 20 timeseries from different sensors/electrical equipment, like temperature, power, current, pressure, flow etc. Here you can imagine being able to represent stuff like that the temperature transmitter 13PT1012 can be used to represent the inletPressure of the pump. Such a generic mechanism would be very helpful in order to represent the connectiveness of things in the process plant, both for maintenance, production optimization and barrier management.
It is becoming clear that Cognite UI/UX is not intuitive for a global scale of sites. How do I not create a laundry list of data sets in the drop down below for each sites? Is there a way to select sites and create a navigation to the right data view of that site and a global view...these challenges remain from a usability perspective.To be honest - CDF just feels like a data dump.
Currently DM instance limit as default is 5 million.A lot of operational scenario may exceed this limit and then a customer or admin from customer side ends up raising a ticket. This is very reactive and by the time an instance limit is increased , it may be 2-3 days to a week. This is basically non productive time and can impact operational activities. I am pretty sure for internal SREs there must be a notification but there should be a way to notify an admin or a group admins to get a notification when certain limits (in this case DM instance) is say 80% of the set limit. This way once the notification is received, a ticket with engineer can be created and before the original limit is reached, a new and higher limit can already be set. This will also give engineering some time to evaluate the request.
It would be beneficial to implement IP access list restrictions for Cognite projects/accounts. This would enable configuration of IP access lists to ensure that users can only connect through secure, predefined corporate networks.For instance, users could be required to connect via the corporate network or an authorized internal VPN. Remote or traveling users could utilize the VPN to securely access the network. If an attempt is made to connect from an unsecured location, such as a public Wi-Fi in a coffee shop, access would be automatically blocked, enhancing security
Currently, it's not possible to use the aggregate query options on both Graphql and SDK for timestamp typed properties. So, for example, if I have a data structure that contains user information and the timestamps of when the user has accessed an application, and I want to get the latest (max), I can't without actually retrieving all the logs and then getting the maximum.
Elkem ASA is no longer going to use pure SQL due to a risk SQL poses. They want to use OpenID/OAuth authentication against DB extractor for this reason.As per my knowledge, we currently don’t have the option to use OAuth2/OpenID authentication with DB extractor.
From the meeting 10/22/24It seems that key matching might be more useful instead of semantics. When searching on the tag, a - was needed to get more results
Transferring asset hierarchy data between the root assets within Cognite poses a challenge due to a limitation within the Cognite Data Fusion (CDF) platform. However, this functionality is readily supported within SAP. In other words, while SAP facilitates the seamless movement of asset hierarchy data across various root assets, Cognite's CDF platform lacks this capability, presenting a constraint in data engineering management and pipeline failures. This limitation results in the real time failures as SAP changes are typically not in our control and business teams perform these changes periodically.
For timeseries, I can only see the datapoints itself, not details and metadata. Not urgent to have this and I understand that IndustrialCanvas is still under development.
The Cognite Pi extractor is currently unable to fill gaps. The only way to force the extractor to re-pull gaps is to:turn the extractor off change the state table (update first = last) change extractor config to enable backfill turn extractor onThis isn't user friendly nor is it documented. It also usually requires an IT admin with access to the extractor's VM and a person who knows where it was installed. It would be amazing to get a no-code feature to do this same thing. Here are two use cases:As an SME i notice that I'm missing data points for a failure or process. I check my source system to verify that the points are missing in CDF. I go to the Extraction Pipeline that monitors this extractor and I enter a timeframe and tag that I wish to gap-fill. As an SME my use case has changed to require more historical data. I want to pull another year's worth of data for a subset of all my tags. I enter a list of tags and set it to backfill just those tags.
BackgroundCurrently CDF allowsuploading a remote extraction config file via python sdk/api - for deployment engineers editing in the UI - for Fusion UI usersWe have noticed in several projects some extraction config files become huge (> 5000 lines). The UI becomes unresponsive in handling such large files. Technical users like deployment engineers are able to edit/update this config through their code but it becomes hard of non-technical users to update stuff in config files (like adding new queries to the ODBC extractor config file). RequirementIf we allow users with a download and an upload option in the UI, it will be helpful for the users to edit/update large config files within the UI. Users can use their local file editor which would be easier to handle.
We have recently made a change to one of source systems which caused a lot of assets to be recreated in CDF with new names, ids, etc. And the old assets were deleted. I have now discovered that our annotated P&ID documents still link to these deleted assets. Is this the desired behavior you’ve intended. From my perspective I would have wanted some clearer indication that the annotation was now invalid or that it was simply deleted when the underlying resource was deleted. An annotation with a dead link is quite useless and frankly just in the way.Let’s say there are integrations reading these file and it’s annotations they would expect the linked item to exist, so if they have not implemented some error handling then the integrations would fail. It is simply to say to just implement the error handling, but it is better (though admittedly harder) to eliminate the issue at the source.Has there been some considerations on this issue? Regards,Markus PettersenAker BP - Data Delivery - CDF Tech Lead
Hello,We want to use the Change Feed from the Cosmos. And this suppport is available in only their Java or .Net sdks.Now when we want to use one of those sdks and now see on Cognite side we see that these both sdks of Cognite are not supported by Cognite directly. So we are in problematic situation here.And hence we request to consider atleast supporting of basic things on Java or .Net Cognite sdk.Also can you clearly list out some risks for continuing usage of java cognite sdk for the time being till these sdks are only community supported ? I was not able to think of anything other than security vulnerabilities fixes.Regards,Neeraj and Tausif
A customer has asked for the xECM for OpenText file extractor
Check the
documentation
Ask the
Community
Take a look
at
Academy
Cognite
Status
Page
Contact
Cognite Support
