agentleFS
PatrickGallucciPatrickGallucci/fabric-skills/skills/fabric-rest-api-remediate/SKILL.md

fabric-rest-api-remediate

Skill17 starsChanged 9 months agoMITView on GitHub

Diagnose and resolve Microsoft Fabric REST API errors including HTTP 401 Unauthorized, 403 Forbidden, 404 Not Found, 429 Throttling, and 5xx server errors. Use when debugging Fabric API authentication failures, Entra ID token issues, insufficient scopes, throttling/rate limiting, long running operation (LRO) polling failures, pagination problems, service principal configuration, or capacity API errors. Covers PowerShell and C# remediate workflows against api.fabric.microsoft.com.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 5 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Start here. Match your HTTP status code to the appropriate workflow.

    line 32
  2. Very specific:

    Every Fabric API error returns an ErrorResponse object.

    line 46
  3. Very specific:

    Never parse message programmatically (may change over time)

    line 63
  4. Very specific:

    Always log requestId for support cases

    line 64
  5. Very specific:

    Check relatedResource to identify which resource caused the error

    line 65
  6. Distinctive:

    Run the diagnostic script to test connectivity:

    line 82
  7. Distinctive:

    If the API call relies on downstream items that don't support service principals, the call fails even if the top-level API does.

    line 146
  8. Distinctive:

    Use the retry helper script:

    line 189
  9. Distinctive:

    Wait and retry later

    line 220
  10. Very specific:

    Use the Location header URL directly (preferred)

    line 260
  11. Very specific:

    If the operation has a result, append /result to the operation URL: GET https://api.fabric.microsoft.com/v1/operations/{operationId}/result

    line 273

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Microsoft Fabric REST API remediate

Structured diagnostic workflows for identifying and resolving errors when calling Microsoft Fabric REST APIs (https://api.fabric.microsoft.com/v1/).

When to Use This Skill

  • Fabric REST API returns an HTTP error (401, 403, 404, 429, 5xx)
  • Authentication or token acquisition fails with MSAL
  • Service principal or managed identity calls are rejected
  • API calls are being throttled (HTTP 429)
  • Long running operations (LRO) fail or time out during polling
  • Pagination with continuationToken stops working or returns incomplete results
  • Capacity API operations fail (create, update, resume, suspend)
  • Need to implement retry logic with exponential backoff for Fabric APIs
  • Debugging errorCode values in Fabric API error responses

Prerequisites

  • Microsoft Entra ID app registration with appropriate Fabric API permissions
  • PowerShell 7+ with Az.Accounts module or MSAL.PS for token acquisition
  • Access to a Microsoft Fabric workspace (Contributor or higher)
  • Tenant setting Service principals can use Fabric APIs enabled (if using service principals)

Quick Diagnostic: Identify Your Error

Start here. Match your HTTP status code to the appropriate workflow.

HTTP StatusError CategoryJump To
401Authentication / TokenAuthentication Failures
403Permissions / AuthorizationPermission Denied
404Resource Not FoundResource Not Found
429Throttling / Rate LimitThrottling
430Spark Compute LimitCapacity Limits
500/502/503Server ErrorServer Errors
202 then failureLRO Polling IssueLRO remediate

Understanding Fabric Error Responses

Every Fabric API error returns an ErrorResponse object. Always capture the requestId for support escalation.

{
  "errorCode": "TokenExpired",
  "message": "The access token has expired.",
  "moreDetails": [],
  "relatedResource": {
    "resourceId": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
    "resourceType": "workspace"
  },
  "requestId": "yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy"
}

Key rules for error handling:

  1. Parse errorCode for programmatic logic (stable, contract-based)
  2. Never parse message programmatically (may change over time)
  3. Always log requestId for support cases
  4. Check relatedResource to identify which resource caused the error

Workflow 1: Authentication Failures (401)

A 401 means the request failed during authentication or access token validation.

Step 1 — Check the errorCode in the response body

errorCodeRoot CauseFix
TokenExpiredAccess token has expiredAcquire a fresh token and retry
InsufficientScopesToken missing required scopesRequest correct scopes in MSAL call

Step 2 — Validate your token

Run the diagnostic script to test connectivity:

./scripts/Test-FabricApiConnection.ps1 -TestEndpoint "workspaces"

Or decode and inspect your token manually:

# Decode JWT token payload (PowerShell)
$token = "<your-access-token>"
$payload = $token.Split('.')[1]
$padding = 4 - ($payload.Length % 4)
if ($padding -lt 4) { $payload += '=' * $padding }
$decoded = [System.Text.Encoding]::UTF8.GetString(
    [System.Convert]::FromBase64String($payload)
) | ConvertFrom-Json
# Check expiration
$expiry = [DateTimeOffset]::FromUnixTimeSeconds($decoded.exp).LocalDateTime
Write-Host "Token expires: $expiry"
Write-Host "Scopes (scp): $($decoded.scp)"
Write-Host "Audience (aud): $($decoded.aud)"

Step 3 — Verify scopes match the API

Fabric REST APIs use scopes in the format https://api.fabric.microsoft.com/<Scope>. Common scopes:

ScopePurpose
Workspace.ReadWrite.AllFull workspace CRUD
Item.ReadWrite.AllFull item CRUD
Item.Read.AllRead-only item access
Item.Execute.AllExecute items (notebooks, pipelines)
<ItemType>.ReadWrite.AllItem-specific CRUD (e.g., Notebook.ReadWrite.All)

Step 4 — If using service principal or managed identity

Confirm the Service principals can use Fabric APIs tenant setting is enabled. Not all APIs support service principals — check the individual API reference page for identity support.

See authentication-guide.md for detailed token acquisition examples in PowerShell and C#.


Workflow 2: Permission Denied (403)

A 403 means the caller is authenticated but lacks permissions on the target resource.

Step 1 — Check the errorCode

errorCodeRoot CauseFix
InsufficientPrivilegesMissing workspace/item permissionsRequest access from workspace admin

Step 2 — Verify workspace role

The calling identity needs at least Contributor role for write operations and Viewer for read operations.

Step 3 — Check for API-specific requirements

Some APIs require Fabric Admin permissions. Admin APIs are under /v1/admin/ paths. Confirm the calling identity has the Fabric Administrator role if targeting admin endpoints.

Step 4 — Service principal considerations

If the API call relies on downstream items that don't support service principals, the call fails even if the top-level API does. For example, calling Items - Create Item to create a data warehouse will fail because data warehouses don't support service principal identity.


Workflow 3: Resource Not Found (404)

A 404 means the specified resource doesn't exist or isn't accessible to the caller.

Step 1 — Check the errorCode

errorCodeRoot CauseFix
WorkspaceNotFoundInvalid workspace IDVerify the workspace GUID
EntityNotFoundInvalid resource IDCheck relatedResource in error for details

Step 2 — Validate your IDs

# List all accessible workspaces to find correct ID
$headers = @{ Authorization = "Bearer $token" }
$response = Invoke-RestMethod -Uri "https://api.fabric.microsoft.com/v1/workspaces" `
    -Headers $headers
$response.value | Select-Object id, displayName | Format-Table

Step 3 — Check URL construction

Fabric API base: https://api.fabric.microsoft.com/v1/workspaces/{workspaceId}/items

Common mistakes: mixing up workspaceId with itemId, using display names instead of GUIDs, or targeting the wrong API version.


Workflow 4: Throttling (429)

A 429 means rate limits have been exceeded. Fabric throttles per caller identity per API within one-minute windows.

Step 1 — Read the Retry-After header

The response includes a Retry-After header indicating seconds to wait. Always use this value rather than hardcoded delays.

Step 2 — Implement bounded retry with exponential backoff

Use the retry helper script:

./scripts/Invoke-FabricApiWithRetry.ps1 `
    -Uri "https://api.fabric.microsoft.com/v1/workspaces" `
    -Token $token `
    -MaxRetries 5

Step 3 — Reduce throttling likelihood

StrategyHow
Use batch/bulk operationsPrefer List Items over individual Get Item calls
Cache metadataStore workspace/item IDs locally, refresh periodically
Spread requestsAdd jitter between calls; avoid burst patterns
Use pagination efficientlyFollow continuationToken to completion in a single pass

See error-codes-reference.md for the complete error code table.


Workflow 5: Capacity and Spark Limits (430)

HTTP 430 is Fabric-specific for Spark compute limits.

Error message: [TooManyRequestsForCapacity] This spark job can't be run because you have hit a spark compute or API rate limit.

Resolution options:

  1. Cancel an active Spark job via the Monitoring Hub
  2. Upgrade to a larger capacity SKU (queue limits scale with SKU size)
  3. Wait and retry later
Fabric SKUQueue Limit
F24
F88
F64 / P164
F128 / P2128
F256 / P3256
TrialNo queueing

Workflow 6: Server Errors (5xx)

Step 1 — Capture the requestId from the response body or header. This is essential for Microsoft Support escalation.

Step 2 — Retry with backoff. Transient 500/502/503 errors often resolve on retry.

Step 3 — Check Fabric service health at Microsoft 365 Service Health or Azure Status.

Step 4 — If persistent, open a support ticket including the requestId, timestamp, and full request/response details.


Workflow 7: Long Running Operation Failures

Fabric uses LRO (Long Running Operations) for time-consuming tasks. An LRO returns HTTP 202 with polling headers.

Step 1 — Capture response headers after 202

HeaderPurpose
LocationURL to poll for operation status
x-ms-operation-idOperation ID for constructing status URL
Retry-AfterSeconds to wait before first poll

Step 2 — Poll the operation status

Two approaches:

  1. Use the Location header URL directly (preferred)
  2. Construct: GET https://api.fabric.microsoft.com/v1/operations/{operationId}

Step 3 — Handle terminal states

StatusMeaningAction
RunningStill in progressContinue polling
SucceededCompleted successfullyFetch result if available
FailedOperation failedCheck error details in response

Step 4 — Get the result

If the operation has a result, append /result to the operation URL: GET https://api.fabric.microsoft.com/v1/operations/{operationId}/result

See long-running-operations.md for complete polling implementation patterns.


Workflow 8: Pagination Issues

Fabric uses continuationToken / continuationUri for paginated responses.

Step 1 — Follow the token to completion

$allItems = @()
$uri = "https://api.fabric.microsoft.com/v1/workspaces/{workspaceId}/items"
do {
    $response = Invoke-RestMethod -Uri $uri -Headers $headers
    $allItems += $response.value
    $uri = $response.continuationUri
} while ($uri)

Step 2 — Common pitfalls

ProblemCauseFix
Missing itemsStopped before token was nullLoop until continuationToken is null or absent
URL encoding errorsToken not URL-encodedUse continuationUri directly (pre-encoded)
Stale tokenToken expired mid-paginationRefresh the bearer token between pages

remediate Decision Tree

API call fails
├── Got HTTP response?
│   ├── 401 → Check token expiry and scopes
│   ├── 403 → Check workspace role and API identity support
│   ├── 404 → Verify workspace/item GUIDs
│   ├── 429 → Honor Retry-After, implement backoff
│   ├── 430 → Spark capacity limit, scale up or wait
│   ├── 5xx → Retry with backoff, check service health
│   └── 202 → LRO started, poll Location header
├── Network error / timeout?
│   ├── Check DNS resolution for api.fabric.microsoft.com
│   ├── Verify no proxy/firewall blocking HTTPS
│   └── Test with: Invoke-RestMethod -Uri "https://api.fabric.microsoft.com/v1/"
└── Token acquisition failed?
    ├── Check Entra app registration exists
    ├── Verify redirect URI matches
    ├── Confirm API permissions granted and admin-consented
    └── See authentication-guide.md

References

  • Error Codes Reference — Complete Fabric API error code table
  • Authentication Guide — Token acquisition patterns for PowerShell and C#
  • Long Running Operations — LRO polling implementation patterns
  • Test-FabricApiConnection.ps1 — Diagnostic connectivity test
  • Invoke-FabricApiWithRetry.ps1 — Retry logic with exponential backoff
  • Fabric-Api-ErrorHandler.template.ps1 — Starter error handling scaffold

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.