Community Discussions

Hey!In our tooling for Power Analysts we compute synthetic-timeseries for them to evaluate scenarios of flow exceeding a threshold  value.     In the current implementation of  SyntheticTimeseries only Average and Interpolation aggregates are allowed. This leads to scenarios where zoomed out (and down-sampled) views of data computed through the SyntheticTimeseries API displays non-informative values. Consider the example below, where the top image is the un-aggregated addition of two timeseries, and the bottom is with the use of SyntheticTimeseries.  

Hi Community!Check out a recent interview with Cognite’s CTO, @Geir Engdahl, discussing the potential implications of generative AI on asset performance management (APM). The interview covers what ChatGPT is and how it works, its potential human and knowledge worker implications, and what to expect from ChatGPT in the APM domain. Our opinion: generative AI is the breakthrough technology the industry needs to deliver the “iPhone moment.”What do you think? Does ChatGPT have the potential to transform legacy asset performance management processes, and if so, how? @Eric Stein-Beldring   

Hello Early AdoptersMy name is Arun and i work as a product manager for the monitoring and alerting services in CDF at Cognite. We are working towards building a solution that is industry standard and solves all needs going forward in the space of monitoring and alerting. Although we have a beta version of monitoring in charts available today we need your help. We would like to understand how you are doing monitoring today, go through a use case to understand the monitoring user flow and finally validate the concept that we have today with you.The session should only take about 90 mins and we can always split this up into two sessions if that is better. You will get to provide feedback to us early as well as help us build a solution that can help make alerting and monitoring much easier to use and do for your use cases.If you are interested in sharing your thoughts and feedback with us please feel free to leave a comment on this thread.

Monitoring in ChartsHello early adopters! It has been a bit quiet on this group as we have been working on a new monitoring solution in CDF and i can now say that after a lot of hard work monitoring in charts in now available on select clusters in CDF namely az-eastus-1, westeurope-1 and europe-west1-1 (Coming soon).Features:Alerting and Monitoring is a central element in a DataOps platform. This allows users to proactively discover issues with equipment, systems and processes.Cognite Data Fusion provides the ability to monitor data as a core capability. It gives users an easy way to set up monitoring jobs and perform further analysis through Charts and other components within CDF.We have the following features available now:Ability to set up a monitoring job on a time series within Charts Upper and lower threshold capabilities out of the box Get notified through CDF UI and emails Filter and manage monitoring jobs within the charts Filter and resolve relevant alerts Stable and scalable

Dear Cognite Hub Members,We are pleased to announce the alpha release of the Cognite SAP Extractor! This extractor retrieves data from SAP ERP or SAP S/4HANA and sends it to CDF Raw through OData protocol.Please note that this is an alpha version, which means it is not ready for production use. We are releasing this alpha to the Cognite Hub community for feedback and testing. Your feedback is valuable in shaping the future of this extractor, so we encourage you to give it a try and let us know what you think.To get started, you can download the SAP Extractor from the “Extract Data” page directly from CDF. Please see the documentation for installation and configuration instructions on the “Cognite SAP Extractor (Alpha)” page under “Extract Data”.    We would appreciate any feedback you have, including bug reports, feature requests and general feedback. Please post your feedback and questions in the “Data Ingestion [Early Adopter]” category in Cognite Hub.Note that we are actively workin

This guide will take you through the steps of setting up a data streaming job on the MQTT ingestion service using the Pluto API.We will throughout the guide provide paths and bodies for the HTTP requests you will make, but we also have an OpenAPI spec you can download and import to your tool of choice (for example Postman) to ease writing these requests. Core conceptsThere are three main object types when using the Pluto service: A source, which models a source of data outside of CDF. This will represent your MQTT broker. The source holds data such as the hostname (or adress) of your broker, credentials for authentication, etc. A destination, which tells Pluto where to put the data output - for example CDF events or data points in a CDF time series. The destination also holds the credentials used when writing to CDF, which is going to be a CDF authentication session (more on that later). A job, which describes the actual job for Pluto to do. It links a source to a destination. It t

Cognite RevealWe are being told that our models need to be downsampled which is resulting in better performance in CDF but poor quality on the viewer. We have consistently noticed that we are unable to have one large point cloud together in the viewer and also the viewing quality is very poor compared to what laser scanning vendor solutions show. I need some attention on this from Product?Why should we downsample and decrease the density of the Point Cloud? does not appear to be the right thing to doWhy should we chop up the 3D model in pieces?

We’ve made a design prototype for the planned alpha version of a MQTT solution. This is only the first iteration of design and is not fully developed yet, but we hope you can give us feedback on the design and presented functionalities. This way we get to know what works and what doesn’t.Please have a look at the video of the prototype and give us your thoughts. Any questions and / or feedback is welcome!(This is an internal preview. Please don’t share it outside this community.) If you’re interested in giving feedback in person or testing out the prototype yourself, feel free to contact us through this post to set up a quick chat. It can be about what you think of the feature, what you would like to improve, if you’d like to test out the UI yourself, etc. We hope to hear from you!

Hi, I wrote a short article some time ago about how data-driven solutions are playing a critical role in helping companies meet their sustainability goals. Here are some key highlights from the article: Sustainability is becoming increasingly important for companies, who are turning to data-driven solutions to monitor, report, and reduce their environmental impact. Cognite has encountered many innovative solutions that promote sustainability, such as automating greenhouse gas emissions reporting and using robotics to detect dangerous leaks. A large set of sustainability metrics in industrial settings are best managed by real-time calculations, which need to be packaged as trusted data products for consumption. Diligent and accurate metrics are important, but decisive actions matter most. With a solid data foundation, companies can optimize operations and drive positive impacts in sustainability. Cognite Data Fusion enables companies to deliver positive impact supporting their sustainab

The cognite replicator fails to replicate asset data and the behaviour is inconsistent. It worked yesterday and not working today. Note that all resource types like time series, events and datapoints are replicated consistently. I am getting below error while replicating the asset data.Please see attached screenshot for the successful run of the asset replication yesterdayTraceback (most recent call last):  File "/Users/j.subhash.parandekar/oid-replicator/oid_replicator/cognite_replicate.py", line 84, in <module>    assets.replicate(SOURCE_CLIENT, DEST_CLIENT, config=cognite_config,  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/cognite/replicator/assets.py", line 324, in replicate    src_dst_ids_assets = create_hierarchy(                         ^^^^^^^^^^^^^^^^^  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/cognite/replicator/assets.py", line 220, in create_hierarchy    updated_assets = replication

In this post we’ll share with you our thoughts on key areas to focus on to empower you as Domain Experts - frontline workers - to efficiently utilize data and analytics in day-to-day operations. Let us know in the comments what you think!  The direction for Cognite Data FusionCognite Data Fusion (CDF) is an Industrial Data platform, and part of our mission is to ensure that we provide a “batteries included” experience to quickly realize value across our target industries Manufacturing, Energy and Power and Renewables.We want CDF to be the go-to tool for not only Data Scientist and deeply technical roles, but also the Domain Experts that know first hand what the operational challenges are and what the optimization potential is. Over the last couple of months, and in our upcoming releases, you will see a focus on enabling you as a Domain Expert to solve industrial data problems through low-code user interfaces - without assistance from “a coder”.  What sort of industrial data problems ar

User sessions are managed via your IDP. Access token lifetime can vary from 60 to 90 minutes.So once the session expires, the user would normally have to sign-in again. How to overcome this?With the OIDC workflow, it is possible to retrieve a new access token without prompting the user to provide credentials again. This is done by finding a valid access token from cache or by finding a valid refresh token from cache and then automatically use it to redeem a new access token. The diagram below shows the normal OIDC workflow: Below you can find a sample code snippet which uses the acquire_token_silent method available through the class: msal.PublicClientApplication:def authenticate_azure(app): accounts = app.get_accounts() if accounts: print("Taking the token silently") creds = app.acquire_token_silent(SCOPES, account=accounts[0]) else: print("Taking token interactively") creds = app.acquire_token_interactive(scopes=SCOPES, port=PORT) return credsY

A common setup in CDF (Cognite Data Fusion) is to first build an asset hierarchy, and then to attach events, timeseries etc. to those assets. One way to do it is to use transformations. In this short article, we’ll take events as an example. It works the same way for timeseries and sequences. Let’s assume we already have an asset hierarchy and we want to attach events to it. We’ll suppose that we have a RAW table, from which we want to create events, that looks like this:   In transformations, to attach an asset to a created event, you need to specify the corresponding asset ID (assetId in the target schema). Asset IDs being automatically generated, we usually prefer using external IDs because we know what they are made of: an ID from an ERP, a value following a naming convention, etc.  Since the asset ID is the expected value for the assetId field in the target schema, we need to retrieve for each asset its ID, based on its external ID. As mentioned in the documentation, we can read f

The coordinates of a document in CDF are normalized. In a normalized document, it has (0,0) coordinates in the upper left corner and (1,1) coordinates in the right bottom corner.Convert xml coordinates to normalized coordinatesTo convert xml coordinates to normalized coordinates you need to divide X coordinates with the width of the page and the Y coordinates with the height of the page. If the xml Y coordinate is 0 at the bottom of the page and the <height> is at the top of the page, the Y coordinate should be flipped as below.Normalized Y = 1- Normalized YOnce the coordinates are flipped;NewMinY = 1 - OldMaxYNewMaxY = 1 - OldMinY Example scenarioIf a user would like to add manual annotations to a P&ID file in CDF the user should convert the xml coordinates of the file to normalized coordinates. 

Hello from the Solutions Portal team!We have a new release coming, which is by far our largest release yet! We hope this release will make it easier toBrowse through your data inside your CDF Project Enable new solutions built on top of CDF Enable new no code solution possibilities  We have a new home screen with new functionality!Any user can now ‘Browse solutions’. Through here, you can view all solutions that Cognite has to offer. If they’re not enabled, a demo can be requested. If you are an admin, SOME applications can be installed through a wizard for free (today, blueprint) and some are automatically installed (Charts).An explorer designed for SMEs. We’ve taken the old ADI explorer, made some improvements based on feedback, and now have an early version out for early adopters.An early version of global search is included also - easily search your way through any CDF resource (with more resources to come!) New CDF application: Blueprint!Blueprint is a no code solution that allows

Digitalization PoCs are commonplace. Real return on investment (ROI) isn’t.So how do investments in digital transformation efforts translate into real value for your business (or company)?We, at Cognite, commissioned a study from Forrester Consulting to examine the potential ROI and business benefits asset-heavy industrial organizations can expect from deploying Cognite Data Fusion. Forrester interviewed six representative customers across our customer base in Oil and Gas, Manufacturing, and Power & Cleantech with experience using Cognite Data Fusion, and found $21.6 million in added net present value at a 400% ROI. Key results of the ROI study include:* $9 million gained through the optimization of heavy machinery and industrial processes* $5.1 million saved through optimized energy use and reduced operational costs* $4.8 million added value through reduced shutdown time* $4.3 million cost reduction by optimizing planned maintenance  Do you want to know what stands behind these nu

I have set up both the environment variables SOURCE_CLIENT_SECRET (Note that the Client secret is generated from the OID widget as per the documentation)and DEST_CLIENT_SECRET and running the replicator with the config file using option 5.I am getting error as per below logAMAC02Z3123LVCJ:oid-replicator j.subhash.parandekar$ poetry run python3 ./oid_replicator/replicate.py 2023-02-17 18:46:48,239 root INFO - Config file - Repeat line 5: 2023-02-17 18:46:48,239 root INFO - Config file - Repeat line 14: 2023-02-17 18:46:48,239 root INFO - Config file - Repeat line 23: Starting replication of resourcesReplicating assets...Traceback (most recent call last):  File "/Users/j.subhash.parandekar/Library/Caches/pypoetry/virtualenvs/oid-replicator-cwaK-6ym-py3.11/lib/python3.11/site-packages/cognite/client/credentials.py", line 364, in _refresh_access_token    token_result = self.__oauth.fetch_token(                   ^^^^^^^^^^^^^^^^^^^^^^^^^  File "/Users/j.subhash.parandekar/Library/Caches/py

I am running the notebook for my own Cognite project as per the example https://github.com/cognitedata/dshub-tutorials/blob/master/advanced/Comparing%20Entity%20Matching%20models%20with%20SDK%20demo.ipynbI am getting below error ModelFailedException: EntityMatchingModel 3531921825900910 failed with error 'JobFailedException: AttributeError('TfidfVectorizer' object has no attribute 'get_feature_names')' at line  results = model.predict(sources=time_series_test, targets=assets).result

Hi, please help me to clarify on the below details as our SLB team is aiming to setup a project with CDF to do some features,Sandbox tenant for SLB  How to connect to production data Embed Python preprocessing script into the Cognite pipeline

Hi, I am trying to transform data from Raw to timeseries in CDF. I made a query and everything seems ok, but when I use distinct to check the external id uniqueness, I get error message:"Cannot have map type columns in DataFrame which calls set operations(intersect, except, etc.), but the type of column metadata is map<string,string>; I don’t know why I get this message, Isn’t it possible to use distinct when wi do MAP or To_Metadata mapping? 

I am the Head of the Services delivery team in Cognite. The team I lead develops and delivers domain use cases based in CDF to our customers in O&G, manufacturing and renewables. I am going to focus here on what are the consequences of corrosion of industrial assets on the economy and the environment, as recently covered in a series of publications (see references section). The economic and environmental cost of corrosion is a significant issue and is much more severe than one could think. While several calculations methods have been used in various studies it is generally accepted that the direct costs of corrosion to the economy are equivalent to roughly 3–4% of a country’s gross domestic product (GDP) [1,2]. The main cost impact is due to the replacement of corroded steel in failing infrastructures (roads, railways, etc.) industrial equipment and buildings. As well, it has been qualitatively estimated that between 25 and 33% of the annual steel production is destroyed once in se

As communicated in this post in December, Cognite Charts is now integrated as a part of fusion.cognite.com. We are planning to retire the stand-alone url charts.cogniteapp.com to have a uniform product experience and single login. Starting February 7th 2023 you will find a message if you access the stand-alone url with instructions on how to login to fusion.cognite.com. You fill find all your charts and data intact, and Charts working as you are used to. There will be a grace period for 2 weeks after February 7th giving room for users to still access their charts from the stand-alone url if you need support to login to fusion.cognite.com.  For those of you who are linking to Charts in your third party applications, Microsoft Teams, etc., this will still function as expected. We will be working on an automatic redirect for coming releases, however still recommend you update the hyperlinks to point to your-organization-name.fusion.cognite.com already now. If you have questions on this, d

When working with CDF or Azure, sometimes I got some errors that were hard to figure out. Experience no doubt will help me figure them out faster, nonetheless this is an important part in enabling usability.

Hey community we are trying to be more active in this cognite hub page and add content consistently from a month to month basis. This will include a series of articles on alerting and monitoring as well as plans regarding where AIR is going into the future.Why Alerting?Automated alerting is an essential part of monitoring. They allow you to spot issues with equipment groups, time series, data quality, pipelines etc.But alerts aren’t always as effective as they could be. In particular, real problems are often lost in a sea of noisy alarms. In short:Alert liberally meaning its ok to spam users rather than not alert at all Make sure that the user has the complete control of how and when they want to be alertedInherent challenges in AlertingSensitivity: Overly sensitive systems cause excessive false positive alerts, while less sensitive systems can miss issues and have false negatives. Determining the correct alerting threshold requires ongoing tuning and refinement. Fatigue: The common ap

 I faced some isuees regarding connection with Azure when working with CDF. I believe there is an area of opportunity there!