rss.livelink.threads-in-node

Pipeline failures: retries, On fail paths or child pipelines?

jubinsoni — Fri, 13 Nov 2026 00:53:43 GMT

I set 3 retries with a 30 second interval on each Copy activity, and the alert fires when the whole pipeline fails. A transient error on one activity can still fail a long run. I tried adding an On fail path for a cleanup step, but I am not sure how the overall pipeline status is reported when On fail or On completion paths are used. What do you rely on: retries, On fail paths, or wrapping risky steps in child pipelines?

Metadata-driven copy pipeline or one pipeline per table: what held up for you?

jubinsoni — Wed, 11 Nov 2026 18:47:42 GMT

Hi all,

We have roughly 80 source tables to ingest into a lakehouse. I am torn between one generic pipeline driven by a control table (ForEach over a Lookup, with the table name, load type and so on as parameters) and a separate pipeline per source.

The generic version is easier to maintain, but I worry about a few things: one bad table slowing or failing the whole run, per-table logging, and the odd source that needs special handling.

If you went generic, where do you keep the control table, how do you cap ForEach concurrency and handle retries, and which sources ended up as exceptions with their own pipeline? And if you went the other way, did the maintenance get painful?

Thanks!

What are the best practices for building reliable data pipelines in Microsoft Fabric Data Factory?

codeautomation — Wed, 11 Nov 2026 17:53:17 GMT

I'm exploring Microsoft Fabric Data Factory and would like to learn how other professionals design and manage reliable data integration pipelines.

When working with multiple data sources and scheduled workflows, maintaining pipeline reliability, handling failures, and monitoring execution are important.

I'd appreciate practical insights on the following:

What are the best practices for designing reusable pipelines in Fabric Data Factory?
How do you handle pipeline failures and configure retry mechanisms?
What is the recommended approach for managing dependencies between pipeline activities?
How can parameters and dynamic expressions make pipelines more flexible?
Which monitoring techniques help identify performance bottlenecks and failed activities?

I'd love to hear about your real-world experiences and recommendations for building scalable and maintainable data pipelines.

Managing Configurations and Secrets Across Fabric Airflow Environments

codeautomation — Mon, 09 Nov 2026 17:32:36 GMT

As Apache Airflow workflows move from development to testing and production, managing environment-specific configurations can become challenging.

For example, a DAG may work perfectly in development but require different connection settings, credentials, storage locations, and environment variables when deployed to production.

I'm interested in understanding how teams handle this in Microsoft Fabric Apache Airflow Jobs.

A few questions for the community:

How do you manage connections and credentials without hardcoding sensitive information in DAG files?
Do you maintain separate configurations for development, staging, and production?
How do you handle environment variables and secret rotation?
What is your preferred approach for promoting DAG changes across environments without manually updating every configuration?

For teams managing multiple Airflow workflows, what practices have helped you make deployments more secure, consistent, and maintainable?

Would love to hear practical experiences and recommendations from the community.

[Open Mirroring] Incremental CSV Upload Overwrites Existing File and Changes Are Not Replicated

dzfamy — Mon, 09 Nov 2026 07:44:49 GMT

Hi Team,

I'd like to report a possible issue with Microsoft Fabric Open Mirroring (manual CSV upload).

A customer reported that the initial CSV upload works, but subsequent uploads (UPDATE/INSERT) are not reflected in the mirrored table.

I was able to reproduce the same behavior in my own Fabric test environment.

Steps to reproduce:

Create a new Open Mirrored Database.
Upload 00000000000000000001.csv with 3 records and create a table with a primary key.
Confirm that all 3 records are successfully replicated.
Upload a second file, 00000000000000000002.csv, using Upload files → Update existing table.
Test both UPDATE (__rowMarker__ = 1) and INSERT (__rowMarker__ = 0) operations.

Actual behavior:

Initial replication works correctly (3 rows).
Subsequent UPDATE/INSERT operations are not reflected.
Replication status remains Running, but the replicated row count stays at 3.
Workspace Monitoring logs show the initial Snapshotting and StartReplicating operations, but no subsequent incremental replication activity.
I checked the Landing Zone using Azure Storage Explorer and confirmed that the second CSV file was not created. Instead, the first CSV file (000...001.csv) was overwritten with the content of the second CSV, despite uploading a different filename (000...002.csv).
The same behavior was reproduced across 3 different tables.

Expected behavior:

The incremental CSV should be processed as a new change file, and the changes should be reflected in the mirrored table.

According to the Microsoft documentation, manual incremental uploads are supported:

https://learn.microsoft.com/en-us/fabric/mirroring/open-mirroring-tutorial

Questions:

Is this a known issue with the Fabric portal's Upload files feature?
Is there any additional configuration required for incremental CSV uploads?
Is there a recommended workaround?

Any guidance from the Fabric Mirroring team would be greatly appreciated.

Thanks!

Pipeline alerting: how do you notify the right team without flooding a channel?

jubinsoni — Sun, 08 Nov 2026 00:49:16 GMT

Hi everyone,

We have around fifty scheduled pipelines and the current alerting is a Teams activity on every failure path, which means a single upstream outage produces dozens of messages. We would like something closer to one alert per incident.

One idea is to write a status row at the end of every run and alert from there:

INSERT INTO ops.PipelineRun (PipelineName, RunId, Status, ErrorMessage, FinishedUtc, OwnerTeam) VALUES (@PipelineName, @RunId, @Status, @ErrorMessage, SYSUTCDATETIME(), @OwnerTeam);

What I am trying to decide

Whether to alert from each pipeline, or from one monitoring pipeline that reads the status table on a schedule.
How to group failures that share a root cause, such as a failing source system, into a single message.
How to route to the owning team and avoid repeating the same alert every hour until it is fixed.

If you have a pattern that works for you, including how you handle retries before an alert is raised, I would like to hear it. Thanks!

How Do You Handle DAG Failures and Retries in Microsoft Fabric Apache Airflow Jobs?

codeautomation — Thu, 05 Nov 2026 09:49:31 GMT

Hi everyone,

I'm exploring how teams handle failure recovery and retry strategies when running Apache Airflow jobs in Microsoft Fabric, particularly for production workflows involving Fabric notebooks, Lakehouses, and data transformation tasks.

One challenge with DAG-based orchestration is ensuring that retries don't accidentally duplicate data, repeat expensive processing, or trigger downstream tasks when upstream operations haven't completed successfully.

I'd appreciate insights from developers and data engineers using Fabric Apache Airflow jobs.

A few questions:

Retry Strategies: How do you configure task-level retries and retry delays for transient failures without repeatedly executing tasks that have already completed?
Idempotency: What approaches do you use to prevent duplicate records or repeated updates when a failed DAG task runs again?
Failure Recovery: When a multi-step DAG fails halfway through execution, do you retry individual tasks, rerun the entire DAG, or use checkpoint-based recovery?
Monitoring and Alerts: How are you tracking failed DAG runs and notifying teams about critical execution failures?
Dependency Management: How do you manage dependencies between Fabric notebooks, pipelines, and other tasks to avoid downstream execution before upstream processing is complete?
Production Operations: What patterns have worked best for handling timeouts, backfills, and unexpected infrastructure interruptions?

I'm particularly interested in real-world implementations where Apache Airflow orchestrates multiple Microsoft Fabric workloads.

For production DAGs, do you prefer automatic retries with idempotent tasks or controlled manual recovery for failed workflows?

Would appreciate any architecture recommendations, implementation examples, or lessons learned from production environments.

Thanks!

What Are the Best Practices for Building Reliable and Scalable Data Pipelines?

codeautomation — Tue, 03 Nov 2026 19:35:40 GMT

As organizations integrate data from multiple applications, databases, APIs, and cloud platforms, building reliable data pipelines becomes increasingly important.

A pipeline that works well with small datasets may encounter problems as data volumes grow and workflows become more complex. Failed activities, dependency issues, schema changes, and inconsistent execution can affect downstream processes and reporting.

I'd like to learn how the community handles these challenges in Microsoft Fabric Data Factory.

A few questions:

How do you design pipelines to handle large datasets efficiently?
What is your approach to implementing retries, failure handling, and recovery mechanisms?
How do you manage dependencies between multiple pipelines and activities?
What strategies do you use for incremental data loading and avoiding duplicate records?
How do you monitor pipeline execution and identify failures before they affect downstream systems?

I'd especially appreciate practical examples involving Microsoft Fabric Data Factory, pipeline orchestration, and automated monitoring.

What practices have helped you build more reliable data pipelines?

How can AI agents improve Microsoft Fabric Data Factory pipeline automation?

codeautomation — Tue, 03 Nov 2026 19:32:07 GMT

I am exploring how AI agents can help make data pipeline development and monitoring more efficient in Microsoft Fabric Data Factory.

Some potential use cases:

Automatically creating pipeline workflows based on business requirements
Detecting pipeline failures and suggesting fixes
Optimizing data movement and transformation steps
Monitoring data quality and identifying anomalies
Triggering automated actions through APIs or external tools

I would like to understand how teams are currently combining AI agents with Fabric Data Factory.

Are you using:

AI-assisted pipeline generation?
Automated error analysis and recovery workflows?
Fabric notebooks with AI models for data operations?
Custom agents connected with Data Factory pipelines?

Would appreciate insights, real-world examples, or architecture patterns from developers working with Fabric Data Factory automation.

Fabric Pipeline: Best way to handle incremental loads from multiple source tables?

Ar_Sh — Mon, 02 Nov 2026 20:02:44 GMT

I'm building a Fabric pipeline to load around 40 SQL Server tables into a Lakehouse. Some tables contain millions of records, so full loads are not practical.

Would you recommend using a single metadata control table to manage incremental loads, or maintaining separate watermarks for each table?

Also, how would you handle tables without a LastModifiedDate column?

Would love to hear how others are handling this in production.

Fabric Pipeline: Best way to handle incremental loads from multiple source tables?

Ar_Sh — Mon, 02 Nov 2026 20:02:40 GMT

I'm building a Fabric pipeline to load around 40 SQL Server tables into a Lakehouse. Some tables contain millions of records, so full loads are not practical.

Would you recommend using a single metadata control table to manage incremental loads, or maintaining separate watermarks for each table?

Also, how would you handle tables without a LastModifiedDate column?

Would love to hear how others are handling this in production.

Error handling pattern for ForEach with partial failures: continue and report, or fail fast?

jubinsoni — Fri, 30 Oct 2026 05:28:08 GMT

Hi everyone,

I have a pipeline that loops over a list of tables with ForEach and runs a Copy activity per item. When one table fails, I would like the remaining tables to still load, and to get one consolidated failure report at the end instead of the whole run failing on the first error.

The pattern I am considering:

Wrap the Copy activity so its failure path (the On fail dependency) logs the table name and error message to a control table, and the ForEach item itself finishes as succeeded.
After the loop, a Script or Lookup activity counts rows in the control table for this run, and a Fail activity raises one error with the summary if the count is above zero.

A sketch of the final check:

SELECT COUNT(*) AS failed_tables FROM ctl.load_log WHERE run_id = @{pipeline().RunId} AND status = 'Failed';

Questions:

Is this the usual approach, or is there a built-in way to get partial-success semantics for a ForEach?
How do you make sure the pipeline run status still reflects the failures so alerts fire?
Any gotchas with the sequential versus parallel setting when logging from inside the loop?

Thanks in advance.

Will Invoke Pipeline (Legacy) be deprecated?

benitovbreugel — Thu, 29 Oct 2026 09:30:45 GMT

Is there currently any plan or timeline to deprecate Invoke Pipeline (Legacy) in Microsoft Fabric?

For same-workspace pipeline orchestration, the Legacy activity is still very convenient. The new Invoke Pipeline activity provides additional capabilities, such as cross-workspace invocation and child-pipeline monitoring, but it also requires a Connection object.

Given the performance concerns that have been reported by multiple users, I’m wondering what the recommended approach is:

Is it safe to continue using Invoke Pipeline (Legacy) for now?
Does Microsoft recommend migrating existing pipelines to the new Invoke Pipeline activity?
Is there any indication that the Legacy activity will eventually be deprecated or removed?
Should we already start rebuilding existing pipeline orchestration using the new activity and Connections?

It would be great to understand the intended direction, especially before investing time in rebuilding existing pipeline implementations.

Thanks!

Can parallel pipeline runs upsert into the same Lakehouse table without commit conflicts?

Microfyx — Wed, 28 Oct 2026 07:01:04 GMT

I want to ingest from 10 DB2 databases into a Lakehouse using one generic pipeline, invoked once per DB. They all write into the same Bronze tables (e.g. .raw_<table>). Every row carries a _db_id, so no two runs ever touch the same record.

The Copy activity uses Table action = Upsert with key columns _db_id + the source primary key. The tables are not partitioned or clustered. In DESCRIBE HISTORY the commits show operation = MERGE, engineInfo = datafactorycopy, and operationParameters, operationMetrics, readVersion and isBlindAppend are all NULL.

I would like to run the DBs in parallel, but we are unsure about commit conflicts. The concurrency control article says conflicts depend on the files a MERGE reads, and that the partition column must be a literal in the merge condition. It describes Spark, though, and I could not find anything about the Copy sink.

Questions:

Does the Copy activity upsert read the whole target table during the MERGE, or can it prune files, for example by _db_id?
Can two parallel Copy upserts (different DBs, same table) fail with something like ConcurrentAppendException? Does the sink retry internally?
Would partitioning the table by _db_id help with the Copy sink, even though we cannot set a literal predicate?
Is there a recommended pattern for this, for example append into per-DB staging tables and a single MERGE afterwards?

For now we run the DBs sequentially. Any experience or documentation pointers would be appreciated.

Copy job vs pipeline + watermark table for incremental loads from Azure SQL: production experiences?

jubinsoni — Tue, 27 Oct 2026 19:39:11 GMT

Hi all,

I'm comparing Copy job with a classic pipeline (Lookup + Copy activity + watermark table) for incremental loads from Azure SQL Database into a Lakehouse, for around a few dozen tables.

The pipeline version is the pattern most of us know:

Lookup: SELECT LastWatermark FROM etl.Watermark WHERE TableName = @{item().name} Copy: SELECT * FROM @{item().name} WHERE ModifiedDate > '@{activity('Lookup').output.firstRow.LastWatermark}' Stored proc: UPDATE etl.Watermark SET LastWatermark = <max ModifiedDate copied>

Copy job seems to cover most of this out of the box (incremental column per table, state tracked for you), which is attractive because there is much less to maintain.

Before choosing, I'd like to understand a few things from people using Copy job in production:

Visibility: when a single table fails, how easy is it to see which table and why, and to rerun only that one?
Resetting state: if a table needs a full reload (schema change, data fix), can you reset the watermark for just that table?
Deletes: for sources where rows are deleted, how are you handling it with Copy job, or do you fall back to mirroring / a pipeline for those tables?
CI/CD: does Copy job work cleanly with Git integration and deployment pipelines, including connection changes between Dev and Prod?

Any experience or lessons learned would be really helpful.

Thanks!

Dataflow Gen2 parameter works with current value but fails when passed from Fabric Pipeline

eyeampraveen — Tue, 27 Oct 2026 14:36:11 GMT

I have a Dataflow Gen2 that calls a SOAP API using Web.Contents().

The Dataflow has an APIKey Text parameter, which I use in Power Query:

v_apikey = Text.From(APIKey)

The value is inserted into the SOAP request header/body before making the Web.Contents() call.

The issue:

API key entered as the Dataflow parameter's current value → Works
Exact same API key passed from a Fabric Pipeline parameter → Fails

The failure eventually results in:

Expression.Error:
There weren't enough elements in the enumeration to complete the operation.
The SOAP XML navigation also shows an empty table:
#table({"Name", "Table"}, {})

I'm trying to understand the runtime execution path:

Pipeline parameter

→ Dataflow activity

→ Dataflow parameter

→ Power Query

→ Web.Contents()

Does Dataflow Gen2 apply Pipeline parameter overrides before Power Query/Web.Contents is evaluated, including any metadata evaluation?

Are there known limitations when a runtime Dataflow parameter is used inside the POST body of Web.Contents()?

Also, is there a safe way to verify that the Pipeline-supplied API key reached Power Query correctly without exposing the key itself?

Unable to Submit a Microsoft Fabric Technical Support Ticket: Only 'Add New Contract' Available

ShubhamJain2 — Tue, 27 Oct 2026 10:01:17 GMT

Hello Microsoft Fabric Community,

I am seeking assistance with accessing Microsoft Fabric technical support for an unresolved Dataflow Gen2 authentication issue.

I have an existing discussion covering the complete technical investigation, troubleshooting history and diagnostic results:

Dataflow Gen2 Lakehouse connection fails with AADSTS70008 | Original Discussion

Rather than repeating the entire troubleshooting history, I am opening this focused discussion to seek help with the remaining support-access issue.

1. Brief technical background

My original Dataflow Gen2 connection to LakeHouse_Training in the Nikolai_Schuler_Udemy workspace fails with:

AADSTS70008: The provided authorization code or refresh token has expired due to inactivity.

Attempts to update credentials or create a new connection to the original Lakehouse also result in an Invalid credentials message.

I completed a clean-room diagnostic test using the same Microsoft Entra tenant, organizational account and Fabric trial capacity.

The results were:

Created a new workspace: Fabric_OAuth_Diagnostic.
Created a new Lakehouse: LH_OAuth_Diagnostic.
Created a new Dataflow Gen2: DF_OAuth_Diagnostic.
Successfully established a fresh Organizational account connection.
Successfully imported CSV data and loaded it into a new Lakehouse table.
The Dataflow execution completed in approximately 22 seconds.

However, connecting to the original LakeHouse_Training from the same diagnostic environment continues to fail.

This indicates that the issue appears to be specific to the original Lakehouse or its connection/authentication path. The underlying root cause remains unconfirmed.

2. Microsoft Fabric technical support access issue

I attempted to obtain technical assistance through the official Microsoft Fabric Support portal and the Contact Support option in Power BI/Fabric.

However, the Support plan dropdown only displays:

Add new contract

The Microsoft Fabric Subscription option is missing, and I cannot proceed with submitting a technical support request.

My environment:

Microsoft Fabric trial capacity.
Fabric Free per-user license.
Organizational member account in a Microsoft Entra tenant.
Global Administrator permissions in the Entra tenant.

I do not have a Unified or Premier support contract.

I also previously submitted an Azure Basic Support request for this issue, but it was closed because technical break-fix support is not covered by that plan.

3. Microsoft documentation

According to Microsoft's current documentation, users with a Fabric Free license can create support tickets for Fabric products.

The documentation also states that the support-plan selection can be made using Microsoft Fabric Subscription or a Unified/Premier support plan. The Add new contract option is intended for users who already have a Unified or Premier contract that is not listed.

Documentation:

https://learn.microsoft.com/en-us/power-bi/support/create-support-ticket

Unfortunately, my support form does not display the documented Microsoft Fabric Subscription option.

4. Assistance requested

Could someone from Microsoft Fabric Support or the Community Support team please clarify:

Why is the Microsoft Fabric Subscription option missing from the support-plan dropdown?
Is there any additional prerequisite or account configuration required for a Fabric Free user with an active trial capacity to submit a technical support ticket?
Is there an alternative route to obtain a technical escalation for this Dataflow Gen2/Lakehouse connection issue?

I would particularly appreciate assistance from a Microsoft Fabric employee or someone who can route this to the appropriate Data Factory or Lakehouse connector support team.

I can provide the detailed Entra sign-in logs, trace IDs, correlation IDs, timestamps and other diagnostic information privately if required.

Thank you for your time and assistance.

Fabric Pipeline fails with 'WorkspaceNotFound' on ADLS Gen2/OneLake connection

debiagui — Mon, 26 Oct 2026 12:14:07 GMT

Hi everyone,

I am facing a critical issue with a Fabric pipeline that started last week. I would appreciate any insights or workarounds.

The Context & Trigger Event:

Last week, all stored connections in my Fabric user profile under "Manage connections and gateways"started expiring
I received a warning/banner indicating that I was very close to reaching the tenant limit of 1,000 stored cloud connections.
In parallel to this credential expiration event, a step in one of my pipelines started failing.

The Pipeline Step:

What it does:It connects to an external SFTP server, downloads data, and writes it directly toOneLake Storage using an Azure Data Lake Storage Gen2 (ADLS Gen2) connection.

The Odd Behavior:

If I go to "Manage connections and gateways", the connection shows as Online and healthy.
I have manually re-entered and updated the credentials there, and the UI says everything is OK (connection test passes in the gateway/connection manager).
However, when I run or test the connection directly inside the Pipeline, it immediately fails with the following error:

ADLS Gen2 operation failed for: Operation returned an invalid status code 'NotFound'. Account: 'onelake'. ErrorCode: 'WorkspaceNotFound'. Message: 'Not Found'. TimeStamp: 'Mon, 26 Oct 2026 11:49:43 GMT'.. Operation returned an invalid status code 'NotFound'

What I have tried so far:

Cleared and re-entered credentials in the connection settings.
Deleted unused connections to free up slot space below the 1,000 limit, but the pipeline error persists
It seems like the pipeline is pointing to a cached, corrupted connection ID token that cannot resolve the workspace path anymore, even though the connection is "green" in settings.

Has anyone encountered this WorkspaceNotFound error on >onelake after hitting the connection limit? Is there a way to force the pipeline to bind to a fresh connection instance?

Thanks in advance!

Dataflow Gen2 Lakehouse connection fails with AADSTS70008

ShubhamJain2 — Tue, 20 Oct 2026 13:41:30 GMT

I am learning Microsoft Fabric and preparing for the DP-600 certification. I have spent several days trying to resolve a Dataflow Gen2 Lakehouse authentication failure, and I need help identifying the root cause.

Environment

Microsoft Fabric trial capacity, with a Fabric Free per-user license.
Personal Azure Free account and a newly created Microsoft Entra tenant.
Workspace: Nikolai_Schuler_Udemy.
Lakehouse: LakeHouse_Training.
Dataflow: Dataflow 1.
Authentication: Organizational account, OAuth 2.0.
Privacy level: Organizational.
Gateway: None.

The Entra tenant was originally created through an Azure signup using a personal Gmail account. I created an onmicrosoft.com member account with Global Administrator permissions and use that account for Fabric. The original Gmail identity exists in the tenant as a guest.

Problem

Dataflow Gen2 cannot connect to the Lakehouse. The query editor displays “Credentials are required to connect to the Lakehouse source.”

When creating or updating the connection, Fabric reports:

AADSTS70008: The provided authorization code or refresh token has expired due to inactivity. Send a new interactive authorization request for this user and resource.

HTTP status: 400
Cluster: https://api.powerbi.com

The same failure occurs when editing credentials for an existing connection. The account picker can open, but reauthentication has not restored access. A newly created connection appears in Manage Connections, yet the Dataflow editor still cannot connect.

Troubleshooting completed

Tested Chrome, Edge, and Edge InPrivate.
Retried interactive sign-in using the onmicrosoft.com member account.
Confirmed Microsoft Authenticator MFA is registered.
Tested existing and newly created Lakehouse connections.
Used Edit credentials in Manage Connections and Gateways.
Confirmed the Lakehouse and SQL analytics endpoint are accessible in Fabric.
Enabled GitHub integration and successfully synchronized the workspace to a public GitHub repository. This did not resolve the authentication error.
Checked Entra sign-in logs. A previous Power BI Data Refresh sign-in failed with error 70008, and the credential update continues to fail with the same error.

Support-access problem

Microsoft's current support documentation says Fabric Free users can submit support tickets for Fabric products. However, the Fabric Help and Support form only offers “Add new contract”; “Microsoft Fabric Subscription” is missing.

I submitted an Azure subscription-management request, but it was closed because Basic Support does not cover technical issues. I still have no technical support case.

Questions

Why does a fresh OAuth authorization continue to fail with AADSTS70008 in this new tenant?
Is there a known issue with Dataflow Gen2, Power Query Online, or the Lakehouse connector's OAuth token handling?
Could the tenant identity setup or Fabric trial capacity affect OAuth authorization or Fabric support eligibility?
How can I obtain a proper Fabric technical support escalation as a Fabric Free user?

Could a Microsoft Fabric employee or Data Factory specialist investigate this and advise on the next diagnostic step? I can provide the full trace ID, correlation ID, request ID, tenant details, and screenshots privately if needed.

Session tag restrictions for high concurrency mode in pipelines

AvdBerg — Tue, 20 Oct 2026 07:59:37 GMT

Hi all,

Our pipeline runs 3 notebooks in succession using high concurrency mode for pipelines with a session tag. The run failed with this error:

Failed to run notebook due to invalid request. [Error: Session tag can only contain letters, numbers, and underscores. (Parameter 'HighConcurrencyModeOptions.SessionTag')]

The tag we used was dataverse-ingest, so the hyphen seems to be the cause. We will switch to underscores, but I could not find this restriction in the Microsoft documentation.

Is the allowed character set for session tags documented anywhere? Are there other limitations (length, reserved words, case sensitivity) we should know about?

Thanks!